@fusion.dev/helix-headless 0.32.0 → 0.33.1

package/custom-elements.json

@@ -352,7 +352,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -850,7 +850,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -1394,7 +1394,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -1914,7 +1914,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -2294,6 +2294,26 @@
               "description": "The bubble's point, forwarded from `Tooltip`.",
               "name": "tooltip-arrow"
             },
+            {
+              "description": "The point while the bubble is below the reading, so on its top edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-bottom"
+            },
+            {
+              "description": "The point while the bubble is left of the reading, so on its right edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-left"
+            },
+            {
+              "description": "The point while the bubble is right of the reading, so on its left edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-right"
+            },
+            {
+              "description": "The point while the bubble is above the reading, so on its bottom edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-top"
+            },
+            {
+              "description": "The bubble while it is below the reading, beside `tooltip`.",
+              "name": "tooltip-bottom"
+            },
             {
               "description": "One column's name and value inside the tooltip.",
               "name": "tooltip-cell"
@@ -2302,10 +2322,22 @@
               "description": "The reading's own name inside the tooltip.",
               "name": "tooltip-label"
             },
+            {
+              "description": "The bubble while it is left of the reading, beside `tooltip`.",
+              "name": "tooltip-left"
+            },
             {
               "description": "A column's name inside the tooltip.",
               "name": "tooltip-name"
             },
+            {
+              "description": "The bubble while it is right of the reading, beside `tooltip`.",
+              "name": "tooltip-right"
+            },
+            {
+              "description": "The bubble while it is above the reading, beside `tooltip`.",
+              "name": "tooltip-top"
+            },
             {
               "description": "A column's value inside the tooltip.",
               "name": "tooltip-value"
@@ -2845,7 +2877,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -3378,7 +3410,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -3910,7 +3942,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -4436,7 +4468,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -4854,6 +4886,26 @@
               "description": "The bubble's point, forwarded from `Tooltip`.",
               "name": "tooltip-arrow"
             },
+            {
+              "description": "The point while the bubble is below the reading, so on its top edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-bottom"
+            },
+            {
+              "description": "The point while the bubble is left of the reading, so on its right edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-left"
+            },
+            {
+              "description": "The point while the bubble is right of the reading, so on its left edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-right"
+            },
+            {
+              "description": "The point while the bubble is above the reading, so on its bottom edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-top"
+            },
+            {
+              "description": "The bubble while it is below the reading, beside `tooltip`.",
+              "name": "tooltip-bottom"
+            },
             {
               "description": "One column's name and value inside the tooltip.",
               "name": "tooltip-cell"
@@ -4862,10 +4914,22 @@
               "description": "The reading's own name inside the tooltip.",
               "name": "tooltip-label"
             },
+            {
+              "description": "The bubble while it is left of the reading, beside `tooltip`.",
+              "name": "tooltip-left"
+            },
             {
               "description": "A column's name inside the tooltip.",
               "name": "tooltip-name"
             },
+            {
+              "description": "The bubble while it is right of the reading, beside `tooltip`.",
+              "name": "tooltip-right"
+            },
+            {
+              "description": "The bubble while it is above the reading, beside `tooltip`.",
+              "name": "tooltip-top"
+            },
             {
               "description": "A column's value inside the tooltip.",
               "name": "tooltip-value"
@@ -5397,7 +5461,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -5913,7 +5977,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -6403,7 +6467,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -6860,7 +6924,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -7295,7 +7359,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -8119,7 +8183,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -8705,7 +8769,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -9154,7 +9218,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -9689,7 +9753,7 @@
               "name": "handleIndicatorKeydown",
               "privacy": "private",
               "readonly": true,
-              "description": "The arrows move between indicators, which is the tablist pattern.\n\nOne tab stop for the set and the arrows within it, so a reader does not\nhave to press Tab once per slide to get past a carousel of twelve. Moving\nchanges the slide, because an indicator that could be focused without\nselecting would leave the highlight and the slide disagreeing with nothing\non screen saying which is which."
+              "description": "The arrows move between indicators, which is the tablist pattern.\n\nOne tab stop for the set and the arrows within it, so a reader does not\nhave to press Tab once per slide to get past a carousel of twelve. Moving\nchanges the slide, because an indicator that could be focused without\nselecting would leave the highlight and the slide disagreeing with nothing\non screen saying which is which.\n\nAlong the line, so on a right-to-left page, where the first indicator is on\nthe right, the left arrow is the one that moves on."
             },
             {
               "kind": "method",
@@ -9760,7 +9824,7 @@
                   }
                 }
               ],
-              "description": "Where a slide's leading edge sits, in the scroller's own coordinates.\n\n**Both directions go through here, because they have to agree.** The\noffsets are measured to the track's *border* box and `scrollTop` and\n`scrollLeft` are measured from its content box, so the two differ by the\nbleed: sending the carousel to a slide and reading back which slide it is\non used the same expression, and correcting one of them alone put a drag\none position off what the arrows did.\n\nUnsubtracted, the leading slide's own offset *is* the bleed, so scrolling\nto index 0 put its edge back on the clip edge. Measured on the insights\ncarousel: `scrollLeft` resting at 4 against a 4px bleed, the first card\nshaved, and no scrolling done to cause it."
+              "description": "Where a slide's leading edge sits, in the scroller's own coordinates.\n\n**Both directions go through here, because they have to agree.** The\noffsets are measured to the track's *border* box and `scrollTop` and\n`scrollLeft` are measured from its content box, so the two differ by the\nbleed: sending the carousel to a slide and reading back which slide it is\non used the same expression, and correcting one of them alone put a drag\none position off what the arrows did.\n\nUnsubtracted, the leading slide's own offset *is* the bleed, so scrolling\nto index 0 put its edge back on the clip edge. Measured on the insights\ncarousel: `scrollLeft` resting at 4 against a 4px bleed, the first card\nshaved, and no scrolling done to cause it.\n\n**On a right-to-left track the leading edge is the right one**, and\n`scrollLeft` counts down from zero as the track moves on. Measured from the\nleft, a carousel showing one slide at a time happened to come out right,\neach slide a whole track to the left of the one before it; one showing\nthree put every one of the first few slides at a positive offset the\nscroller cannot reach, so it clamped to zero and the next arrow moved the\ndots and not the slides. Measured in Chromium on the several-across\nexample: three presses of next, `index` at 2 and `scrollLeft` still at 0."
             },
             {
               "kind": "method",
@@ -10012,7 +10076,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -10549,7 +10613,7 @@
                   "text": "TemplateResult"
                 }
               },
-              "description": "The control that offers the turn, behind a seam so a tier can draw its own.\n\n**A glyph with the label as its accessible name, rather than the label as\ntext.** The word \"Send\" beside a field is the widest thing in the row and\nsays nothing a return key does not. The return glyph is the mark because\nEnter is what actually sends, which `submit-on` says out loud: the control\nand the key are one affordance and should look like it. `send-label` is the\naccessible name rather than a visible one, so a screen reader still hears\nthe word."
+              "description": "The control that offers the turn, behind a seam so a tier can draw its own.\n\n**A glyph with the label as its accessible name, rather than the label as\ntext.** The word \"Send\" beside a field is the widest thing in the row and\nsays nothing a return key does not. The return glyph is the mark because\nEnter is what actually sends, which `submit-on` says out loud: the control\nand the key are one affordance and should look like it. `send-label` is the\naccessible name rather than a visible one, so a screen reader still hears\nthe word.\n\n**Not mirrored on a right-to-left page**, where every caret meaning next\nis. It pictures a key rather than a way to go, like the arrows `Kbd`\ndraws, and a picture of a key is not turned round with the text."
             },
             {
               "kind": "field",
@@ -10681,7 +10745,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -11420,7 +11484,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -11773,7 +11837,7 @@
       "declarations": [
         {
           "kind": "class",
-          "description": "**It has no opinion about where it sits.** A chat is asked for as a floating\nbutton, a page, a palette, a sidebar, a modal and an inline block, and none\nof those is a variant of a conversation: they are containers, and the\ncontainers already exist. Nothing here is positioned, nothing has a backdrop,\nand nothing owns a trigger. If a form factor needs a mode on this element,\nthe split is wrong.\n\n**Scroll ownership is the behavior everything else depends on.** A transcript\nthat grows wants to stay at the bottom, except when the reader has scrolled\nup to read, where pinning is hostile. The rule: pin while the reader is\nwithin `threshold` of the bottom, and let go the moment they leave it. A turn\ngrowing under a pinned reader is not the reader leaving, so a streamed answer\nis followed to its last line.\n\n**History loads upward, under a reader who does not move.** With `history`\nset to `more`, reaching the top fires `helix-history`; the consumer sets\n`loading` while it fetches and prepends the older turns, and the turn the\nreader was looking at stays where it was on screen while the room above it\ngrows.\n\n**A turn is announced once, when it settles**, and not by the log. A log\nthat is a live region reads what enters it, which is every token of a stream\nwhere `aria-busy` is not honored, every page of history as it loads, and\nevery window of a long transcript as the reader scrolls through it. So the\nlog is `aria-live=\"off\"`, and a status region of its own says each new turn\nwhen it has finished arriving: never the reader's own words unless they\nfailed to send, never a turn loaded above the ones already there, and never\nwhat was in the transcript when it opened.\n\n**Turns are markup children or data, not both.** Children serve whole from a\nserver and are what most conversations want. `turns` is for the transcript\ntoo long to hold as elements: it is drawn a window at a time, measured as it\ngoes, and the log needs a block size of its own for that to mean anything.",
+          "description": "**It has no opinion about where it sits.** A chat is asked for as a floating\nbutton, a page, a palette, a sidebar, a modal and an inline block, and none\nof those is a variant of a conversation: they are containers, and the\ncontainers already exist. Nothing here is positioned, nothing has a backdrop,\nand nothing owns a trigger. If a form factor needs a mode on this element,\nthe split is wrong.\n\n**Scroll ownership is the behavior everything else depends on.** A transcript\nthat grows wants to stay at the bottom, except when the reader has scrolled\nup to read, where pinning is hostile. The rule: pin while the reader is\nwithin `threshold` of the bottom, and let go the moment they leave it. A turn\ngrowing under a pinned reader is not the reader leaving, so a streamed answer\nis followed to its last line.\n\n**History loads upward, under a reader who does not move.** With `history`\nset to `more`, reaching the top fires `helix-history`; the consumer sets\n`loading` while it fetches and prepends the older turns, and the turn the\nreader was looking at stays where it was on screen while the room above it\ngrows.\n\n**A turn is announced once, when it settles**, and not by the log. A log\nthat is a live region reads what enters it, which is every token of a stream\nwhere `aria-busy` is not honored, every page of history as it loads, and\nevery window of a long transcript as the reader scrolls through it. So the\nlog is `aria-live=\"off\"`, and a status region of its own says each new turn\nwhen it has finished arriving: never the reader's own words unless they\nfailed to send, never a turn loaded above the ones already there, and never\nwhat was in the transcript when it opened.\n\n**Turns are markup children or data, not both.** Children serve whole from a\nserver and are what most conversations want. `turns` is for the transcript\ntoo long to hold as elements: it is drawn a window at a time, measured as it\ngoes, and the log needs a block size of its own for that to mean anything.\n\n**A fed turn is styled through this element, not its own.** Its message is\ndrawn in this shadow root, where `[author=\"me\"]` cannot reach it, so every\nfed turn's parts are forwarded as `message-*`, and the ones that change with\nthe turn a second time while it holds: `-mine` for a turn by `me`, and\n`message-meta-failed`, `message-content-sending` and\n`message-content-streaming` for its state. Every other author has the plain\nnames only, so a conversation styling each participant apart passes its\nturns as children.",
           "name": "HelixHeadlessChat",
           "cssParts": [
             {
@@ -11828,14 +11892,34 @@
               "description": "A fed turn's leading slot's box.",
               "name": "message-avatar"
             },
+            {
+              "description": "The leading slot's box on a fed turn by `me`.",
+              "name": "message-avatar-mine"
+            },
             {
               "description": "What a fed turn said, and the line under it.",
               "name": "message-body"
             },
+            {
+              "description": "What a fed turn by `me` said, and the line under it.",
+              "name": "message-body-mine"
+            },
             {
               "description": "A fed turn's text.",
               "name": "message-content"
             },
+            {
+              "description": "The text of a fed turn by `me`.",
+              "name": "message-content-mine"
+            },
+            {
+              "description": "A fed turn's text while it is `sending`.",
+              "name": "message-content-sending"
+            },
+            {
+              "description": "A fed turn's text while it is `streaming`.",
+              "name": "message-content-streaming"
+            },
             {
               "description": "A fed turn's default wait.",
               "name": "message-indicator"
@@ -11848,6 +11932,18 @@
               "description": "The line under a fed turn: its time, its state, and its controls.",
               "name": "message-meta"
             },
+            {
+              "description": "The line under a fed turn that `failed`, which says so.",
+              "name": "message-meta-failed"
+            },
+            {
+              "description": "The line under a fed turn by `me`.",
+              "name": "message-meta-mine"
+            },
+            {
+              "description": "A fed turn by `me`: the message drawn for it.",
+              "name": "message-mine"
+            },
             {
               "description": "The control offered on a fed turn that failed.",
               "name": "message-retry"
@@ -12736,7 +12832,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -13101,6 +13197,11 @@
                 "description": "A conversation with the four states a turn can be in, and a composer under it.",
                 "layout": "padded",
                 "slug": "preview"
+              },
+              {
+                "description": "Two thousand turns handed over as data, and the reader's own told apart from outside.",
+                "min": 420,
+                "slug": "transcript"
               }
             ],
             "source": "helix/headless/src/chat"
@@ -13769,7 +13870,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -14940,7 +15041,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -15499,7 +15600,7 @@
               "name": "context"
             },
             {
-              "description": "The copy button.",
+              "description": "The copy button: in the header where there is one, and otherwise beside the code.",
               "name": "copy"
             },
             {
@@ -15518,6 +15619,38 @@
               "description": "The point, resting against the trigger.",
               "name": "copy-tooltip-arrow"
             },
+            {
+              "description": "The point while the bubble is below the copy button, so on its top edge, beside `copy-tooltip-arrow`.",
+              "name": "copy-tooltip-arrow-bottom"
+            },
+            {
+              "description": "The point while the bubble is left of the copy button, so on its right edge, beside `copy-tooltip-arrow`.",
+              "name": "copy-tooltip-arrow-left"
+            },
+            {
+              "description": "The point while the bubble is right of the copy button, so on its left edge, beside `copy-tooltip-arrow`.",
+              "name": "copy-tooltip-arrow-right"
+            },
+            {
+              "description": "The point while the bubble is above the copy button, so on its bottom edge, beside `copy-tooltip-arrow`.",
+              "name": "copy-tooltip-arrow-top"
+            },
+            {
+              "description": "The bubble while it is below the copy button, beside `copy-tooltip`.",
+              "name": "copy-tooltip-bottom"
+            },
+            {
+              "description": "The bubble while it is left of the copy button, beside `copy-tooltip`.",
+              "name": "copy-tooltip-left"
+            },
+            {
+              "description": "The bubble while it is right of the copy button, beside `copy-tooltip`.",
+              "name": "copy-tooltip-right"
+            },
+            {
+              "description": "The bubble while it is above the copy button, beside `copy-tooltip`.",
+              "name": "copy-tooltip-top"
+            },
             {
               "description": "In a diff, one file's header and block.",
               "name": "file"
@@ -15527,11 +15660,11 @@
               "name": "filler"
             },
             {
-              "description": "The code and the copy button beside it.",
+              "description": "The code, and the copy button beside it where there is no header.",
               "name": "frame"
             },
             {
-              "description": "The bar carrying the name, where there is one.",
+              "description": "The bar carrying the name and the copy button, where there is a name.",
               "name": "header"
             },
             {
@@ -15650,6 +15783,51 @@
               "description": "The code as text, read from what the default slot is given rather than\nstored.\n\n**The host's own children rather than the slot's assigned nodes**, which\nare the same nodes: text, and every element without a `slot`. A slot exists\nonly once a render has made it, so reading through one left the first render\nwith nothing to copy.\n\nClient only, and it has to be: the server has no children to read. The\ncopy button is a client-side affordance anyway, so nothing is lost, and\ntaking the text from a property instead would be a second copy of the code\nthat goes stale the moment the slot changes. Empty while hydrating what a\nserver sent, which rendered it empty.",
               "readonly": true
             },
+            {
+              "kind": "method",
+              "name": "renderHeader",
+              "privacy": "private",
+              "return": {
+                "type": {
+                  "text": "TemplateResult | typeof nothing"
+                }
+              },
+              "parameters": [
+                {
+                  "name": "name",
+                  "type": {
+                    "text": "string"
+                  }
+                },
+                {
+                  "name": "value",
+                  "optional": true,
+                  "type": {
+                    "text": "string"
+                  }
+                }
+              ],
+              "description": "The name with the copy button beside it, or nothing when there is no name.\n\n**The button is in the header wherever there is one**, because beside the\ncode it stands over the block and the block has to make room for it. A\ndiff cannot: its washes run to the edge, so the button stood on the hunk\nline and the line under it. In the header it is over nothing.\n\n**No name draws no header**, rather than a bar to hold one button, which\nturned a `Usage` section of one-line imports into a run of empty strips.\nThe button then stands beside the code, in room the block keeps for it."
+            },
+            {
+              "kind": "method",
+              "name": "renderCopy",
+              "privacy": "private",
+              "return": {
+                "type": {
+                  "text": "TemplateResult | typeof nothing"
+                }
+              },
+              "parameters": [
+                {
+                  "name": "value",
+                  "optional": true,
+                  "type": {
+                    "text": "string"
+                  }
+                }
+              ]
+            },
             {
               "kind": "method",
               "name": "renderCopyButton",
@@ -15902,7 +16080,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -16361,7 +16539,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -17023,13 +17201,36 @@
               "name": "panel",
               "privacy": "protected",
               "readonly": true,
-              "default": "new PopoverController(this, { anchor: () => this.partNamed('field'), /* The surface *and* the trigger. * * A control is not always one element: the button that opens this panel is * part of it, and a press on it is not a press outside. Scoped to the * surface alone, `pointerdown` on the trigger closes the panel and the * trigger's own `toggleOpen` reopens it on the `click` that follows, so the * panel can never be closed by the button that opened it. * `DismissController` documents this failure and takes a list for it. */ boundary: (): readonly (Element | null | undefined)[] => [ this.partNamed('surface'), this.partNamed('trigger'), ], /* Wider than the four every other panel here takes. Those sit against a control the same width as themselves and read as an extension of it; this one is a picker of its own size hanging off a field, so the gap is what says the two are separate things rather than one control that grew. */ offset: () => 8, onDismiss: () => this.requestOpen(false), open: () => this.open, placement: 'bottom-start', surface: () => this.partNamed('surface'), })",
+              "default": "new PopoverController(this, { anchor: () => this.partNamed('field'), /* The surface *and* the trigger. * * A control is not always one element: the button that opens this panel is * part of it, and a press on it is not a press outside. Scoped to the * surface alone, `pointerdown` on the trigger closes the panel and the * trigger's own `toggleOpen` reopens it on the `click` that follows, so the * panel can never be closed by the button that opened it. * `DismissController` documents this failure and takes a list for it. */ boundary: (): readonly (Element | null | undefined)[] => [ this.partNamed('surface'), this.partNamed('trigger'), ], /* Wider than the four every other panel here takes. Those sit against a control the same width as themselves and read as an extension of it; this one is a picker of its own size hanging off a field, so the gap is what says the two are separate things rather than one control that grew. */ offset: () => 8, onDismiss: () => this.requestOpen(false), onPlace: (_placement, room) => this.placed(room), open: () => this.open, placement: 'bottom-start', surface: () => this.partNamed('surface'), })",
               "description": "The overlay, placed against the field rather than the trigger.\n\nAgainst the field because that is the box a reader sees as the control: a\ncalendar hung off a small button at one end of a wide field reads as\nbelonging to the button. `Select` anchors to its trigger for the opposite\nreason, its trigger being the whole control.",
               "inheritedFrom": {
                 "name": "MixinPickerField",
                 "module": "src/mixins/mixin-picker-field/mixin-picker-field.ts"
               }
             },
+            {
+              "kind": "method",
+              "name": "placed",
+              "privacy": "protected",
+              "return": {
+                "type": {
+                  "text": "void"
+                }
+              },
+              "parameters": [
+                {
+                  "name": "_room",
+                  "type": {
+                    "text": "number | undefined"
+                  }
+                }
+              ],
+              "description": "Told the panel has been placed, and how wide the window leaves it room to\nbe, in pixels.\n\nNothing, unless a family fits what its panel holds to the room it has.",
+              "inheritedFrom": {
+                "name": "MixinPickerField",
+                "module": "src/mixins/mixin-picker-field/mixin-picker-field.ts"
+              }
+            },
             {
               "kind": "field",
               "name": "control",
@@ -17888,7 +18089,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -19476,7 +19677,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -21407,7 +21608,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -22467,7 +22668,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -22894,6 +23095,38 @@
             {
               "description": "The point, resting against the trigger.",
               "name": "tooltip-arrow"
+            },
+            {
+              "description": "The point while the bubble is below the button, so on its top edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-bottom"
+            },
+            {
+              "description": "The point while the bubble is left of the button, so on its right edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-left"
+            },
+            {
+              "description": "The point while the bubble is right of the button, so on its left edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-right"
+            },
+            {
+              "description": "The point while the bubble is above the button, so on its bottom edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-top"
+            },
+            {
+              "description": "The bubble while it is below the button, beside `tooltip`.",
+              "name": "tooltip-bottom"
+            },
+            {
+              "description": "The bubble while it is left of the button, beside `tooltip`.",
+              "name": "tooltip-left"
+            },
+            {
+              "description": "The bubble while it is right of the button, beside `tooltip`.",
+              "name": "tooltip-right"
+            },
+            {
+              "description": "The bubble while it is above the button, beside `tooltip`.",
+              "name": "tooltip-top"
             }
           ],
           "slots": [
@@ -23593,7 +23826,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -24718,6 +24951,16 @@
               "description": "The columns in the order a reader arranged them.\n\nThe ones with a field are put in `order`'s order, those it does not name\nafter them in the order they were written, and a column with no field stays\nwhere it was written: it has nothing to be named by.",
               "readonly": true
             },
+            {
+              "kind": "field",
+              "name": "shownColumns",
+              "type": {
+                "text": "HelixHeadlessTableColumn[]"
+              },
+              "privacy": "private",
+              "description": "The columns drawn: the arranged ones less any that are hidden.\n\n**A hidden column takes no track and draws no cells.** Left in, its\nheading went and its track and its cell in every row stayed, so the header\nrow came up a heading short, the first cell filled the gap, and every row\nafter it began a column late. It keeps its place in `order` and in the\nsort, which are by field, so showing it again puts it back where it was.",
+              "readonly": true
+            },
             {
               "kind": "field",
               "name": "fields",
@@ -24972,7 +25215,38 @@
                   }
                 }
               ],
-              "description": "Moves a column by some places among the ones that move."
+              "description": "Moves a column by some places among the ones that move and are drawn.\n\nCounted among the drawn ones, so a key moves the heading past one a reader\ncan see: counted among all of them, a step past a hidden column changed the\norder and nothing on screen. A hidden column keeps its place among the rest."
+            },
+            {
+              "kind": "method",
+              "name": "placed",
+              "privacy": "private",
+              "return": {
+                "type": {
+                  "text": "string[]"
+                }
+              },
+              "parameters": [
+                {
+                  "name": "field",
+                  "type": {
+                    "text": "string"
+                  }
+                },
+                {
+                  "name": "beside",
+                  "type": {
+                    "text": "string"
+                  }
+                },
+                {
+                  "name": "after",
+                  "type": {
+                    "text": "boolean"
+                  }
+                }
+              ],
+              "description": "Every field in order, with one taken out and put back beside another."
             },
             {
               "kind": "method",
@@ -24982,7 +25256,7 @@
                   "text": "string"
                 }
               },
-              "description": "The result as CSV: every row the filter lets through, in the order on\nscreen, in the columns on screen, headed by their headings.\n\n**Values, not cells.** A cell may be a template, a badge or a link, so each\ncolumn gives its field's value, which is what a spreadsheet wants anyway.\nA column with no field has no value and is left out. In `server` mode this\nis the page it holds, which is all it can know.\n\n**A text value that a spreadsheet would run is quoted as text.** A cell\nopening with `=`, `+`, `-` or `@` is a formula to a spreadsheet, and a\nresult is often something a reader did not write: those values gain a\nleading apostrophe, the mark spreadsheets read as \"this is text\". Numbers are\nleft alone, so a negative one stays a number.",
+              "description": "The result as CSV: every row the filter lets through, in the order on\nscreen, in the columns on screen, headed by their headings.\n\n**Values, not cells.** A cell may be a template, a badge or a link, so each\ncolumn gives its field's value, which is what a spreadsheet wants anyway.\nA column with no field has no value and is left out, and so is a hidden\none, which is not on screen. In `server` mode this is the page it holds,\nwhich is all it can know.\n\n**A text value that a spreadsheet would run is quoted as text.** A cell\nopening with `=`, `+`, `-` or `@` is a formula to a spreadsheet, and a\nresult is often something a reader did not write: those values gain a\nleading apostrophe, the mark spreadsheets read as \"this is text\". Numbers are\nleft alone, so a negative one stays a number.",
               "type": {
                 "text": "csv() => string"
               }
@@ -25657,7 +25931,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -26771,9 +27045,52 @@
                 "text": "number"
               },
               "default": "1",
-              "description": "How many months the calendar shows, side by side.\n\nWorth more than one mainly for a range, where picking a start in March and\nan end in April otherwise means paging between them and losing sight of the\nfirst. The panel is sized by its content, so this widens it.",
+              "description": "How many months the calendar shows, side by side.\n\nWorth more than one mainly for a range, where picking a start in March and\nan end in April otherwise means paging between them and losing sight of the\nfirst. The panel is sized by its content, so this widens it, and where the\nwindow has no room for that width it shows one month until there is.",
               "attribute": "months"
             },
+            {
+              "kind": "field",
+              "name": "fitted",
+              "type": {
+                "text": "number | undefined"
+              },
+              "privacy": "private",
+              "default": "undefined",
+              "description": "How many months the panel is showing, where that is fewer than `months`.\n\n**One month rather than a squeezed two**, and decided here rather than in a\nstylesheet because what changes is markup. Two months measure 600 pixels\nacross, measured 15 Sep 2026, and in a 390 window that ran past both\nedges. A rule could only clip them, which hides the next-month control with\nthe month it sits on, or stack them, which makes a panel taller than most\nof a phone. The count is behavior, as a carousel's per-view is."
+            },
+            {
+              "kind": "field",
+              "name": "wide",
+              "type": {
+                "text": "number | undefined"
+              },
+              "privacy": "private",
+              "default": "undefined",
+              "description": "How wide the panel was the last time it held every month, in pixels.\n\nMeasured rather than estimated, because a month's width is its locale's:\nthe weekday names set it. Remembered, so a panel showing one month knows\nwhat it would take to show them all, and the answer cannot flip back and\nforth between two measurements of different things."
+            },
+            {
+              "kind": "method",
+              "name": "placed",
+              "privacy": "protected",
+              "return": {
+                "type": {
+                  "text": "void"
+                }
+              },
+              "parameters": [
+                {
+                  "name": "room",
+                  "type": {
+                    "text": "number | undefined"
+                  }
+                }
+              ],
+              "description": "Shows one month where the window has no room for them all, and all of them\nagain once it does.\n\nCalled after every placement, with the room the floating pass measured,\nwhich is the window less a margin on each side. The panel hangs in the top\nlayer, so the window is its box, the argument Dialog makes for reading the\nviewport. The surface's own width is read, since it is sized by its content\nand never capped.",
+              "inheritedFrom": {
+                "name": "MixinPickerField",
+                "module": "src/mixins/mixin-picker-field/mixin-picker-field.ts"
+              }
+            },
             {
               "kind": "field",
               "name": "placeholder",
@@ -27057,7 +27374,7 @@
               "name": "panel",
               "privacy": "protected",
               "readonly": true,
-              "default": "new PopoverController(this, { anchor: () => this.partNamed('field'), /* The surface *and* the trigger. * * A control is not always one element: the button that opens this panel is * part of it, and a press on it is not a press outside. Scoped to the * surface alone, `pointerdown` on the trigger closes the panel and the * trigger's own `toggleOpen` reopens it on the `click` that follows, so the * panel can never be closed by the button that opened it. * `DismissController` documents this failure and takes a list for it. */ boundary: (): readonly (Element | null | undefined)[] => [ this.partNamed('surface'), this.partNamed('trigger'), ], /* Wider than the four every other panel here takes. Those sit against a control the same width as themselves and read as an extension of it; this one is a picker of its own size hanging off a field, so the gap is what says the two are separate things rather than one control that grew. */ offset: () => 8, onDismiss: () => this.requestOpen(false), open: () => this.open, placement: 'bottom-start', surface: () => this.partNamed('surface'), })",
+              "default": "new PopoverController(this, { anchor: () => this.partNamed('field'), /* The surface *and* the trigger. * * A control is not always one element: the button that opens this panel is * part of it, and a press on it is not a press outside. Scoped to the * surface alone, `pointerdown` on the trigger closes the panel and the * trigger's own `toggleOpen` reopens it on the `click` that follows, so the * panel can never be closed by the button that opened it. * `DismissController` documents this failure and takes a list for it. */ boundary: (): readonly (Element | null | undefined)[] => [ this.partNamed('surface'), this.partNamed('trigger'), ], /* Wider than the four every other panel here takes. Those sit against a control the same width as themselves and read as an extension of it; this one is a picker of its own size hanging off a field, so the gap is what says the two are separate things rather than one control that grew. */ offset: () => 8, onDismiss: () => this.requestOpen(false), onPlace: (_placement, room) => this.placed(room), open: () => this.open, placement: 'bottom-start', surface: () => this.partNamed('surface'), })",
               "description": "The overlay, placed against the field rather than the trigger.\n\nAgainst the field because that is the box a reader sees as the control: a\ncalendar hung off a small button at one end of a wide field reads as\nbelonging to the button. `Select` anchors to its trigger for the opposite\nreason, its trigger being the whole control.",
               "inheritedFrom": {
                 "name": "MixinPickerField",
@@ -28106,7 +28423,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -28458,7 +28775,7 @@
                 "text": "number"
               },
               "default": "1",
-              "description": "How many months the calendar shows, side by side.\n\nWorth more than one mainly for a range, where picking a start in March and\nan end in April otherwise means paging between them and losing sight of the\nfirst. The panel is sized by its content, so this widens it.",
+              "description": "How many months the calendar shows, side by side.\n\nWorth more than one mainly for a range, where picking a start in March and\nan end in April otherwise means paging between them and losing sight of the\nfirst. The panel is sized by its content, so this widens it, and where the\nwindow has no room for that width it shows one month until there is.",
               "fieldName": "months"
             },
             {
@@ -29774,7 +30091,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -30466,7 +30783,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -30881,7 +31198,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -31593,7 +31910,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -32149,7 +32466,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -32509,6 +32826,26 @@
               "description": "The bubble's point, forwarded from `Tooltip`.",
               "name": "tooltip-arrow"
             },
+            {
+              "description": "The point while the bubble is below the reading, so on its top edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-bottom"
+            },
+            {
+              "description": "The point while the bubble is left of the reading, so on its right edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-left"
+            },
+            {
+              "description": "The point while the bubble is right of the reading, so on its left edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-right"
+            },
+            {
+              "description": "The point while the bubble is above the reading, so on its bottom edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-top"
+            },
+            {
+              "description": "The bubble while it is below the reading, beside `tooltip`.",
+              "name": "tooltip-bottom"
+            },
             {
               "description": "One column's name and value inside the tooltip.",
               "name": "tooltip-cell"
@@ -32517,10 +32854,22 @@
               "description": "The slice's own name inside the tooltip.",
               "name": "tooltip-label"
             },
+            {
+              "description": "The bubble while it is left of the reading, beside `tooltip`.",
+              "name": "tooltip-left"
+            },
             {
               "description": "A column's name inside the tooltip.",
               "name": "tooltip-name"
             },
+            {
+              "description": "The bubble while it is right of the reading, beside `tooltip`.",
+              "name": "tooltip-right"
+            },
+            {
+              "description": "The bubble while it is above the reading, beside `tooltip`.",
+              "name": "tooltip-top"
+            },
             {
               "description": "A column's value inside the tooltip.",
               "name": "tooltip-value"
@@ -32567,6 +32916,16 @@
               "description": "The fraction of the circle the shares are spread over.",
               "readonly": true
             },
+            {
+              "kind": "field",
+              "name": "origin",
+              "type": {
+                "text": "number"
+              },
+              "privacy": "private",
+              "description": "Where the first slice starts, as an angle from three o'clock.\n\nTwelve for a whole ring and nine for a half, the turns the stylesheet gives\nthe circles. Anything drawn or placed by arithmetic has to start from the\nsame place, or it describes a slice the ring is not showing there.",
+              "readonly": true
+            },
             {
               "kind": "field",
               "name": "thickness",
@@ -32673,7 +33032,7 @@
                   }
                 }
               ],
-              "description": "The same slice again, as a shape with an inside.\n\n**A stroked arc has no boundary to draw, which is why this exists.** The\nsegment above is a dash on a circle: the mark *is* the edge, so there is\nnothing for an outline to trace and nothing within it to fade. Anything\nwanting those two separately needs the slice as a closed path, and this is\nit: the outer arc, a step inward, the inner arc back, closed.\n\n**Rendered always and painted only sometimes**, which is what keeps this on\nthis tier. Whether a boundary is drawn is a styled question and the answer\nlives in a stylesheet; whether the geometry a boundary would need exists at\nall is structural, and a component cannot conjure a shape from CSS. So the\npath is here for every donut and is `fill: none; stroke: none` until\nsomething asks for it.\n\n**Not a replacement for the dash.** The dash carries three jobs and only\none of them is the shape: a share growing is a `stroke-dasharray`\ninterpolating with no reflow, and `@starting-style` on that same property is\nthe whole reason the ring sweeps in rather than appearing. A `d` cannot do\neither. CSS interpolates one only between paths of identical command\nstructure, and the large-arc flag flips the moment a slice crosses a half\nturn, so a growing slice would pop partway through the transition it was\nmeant to smooth. Both animations stay on the dash and this path follows it.\n\nThe turn to twelve o'clock is arithmetic here rather than the rotation the\nstylesheet gives the circles. That rule uses `transform-box: fill-box`, so\nit spins each element about its own bounding box, which is the ring's center\nfor a circle and is not for a sector."
+              "description": "The same slice again, as a shape with an inside.\n\n**A stroked arc has no boundary to draw, which is why this exists.** The\nsegment above is a dash on a circle: the mark *is* the edge, so there is\nnothing for an outline to trace and nothing within it to fade. Anything\nwanting those two separately needs the slice as a closed path, and this is\nit: the outer arc, a step inward, the inner arc back, closed.\n\n**Rendered always and painted only sometimes**, which is what keeps this on\nthis tier. Whether a boundary is drawn is a styled question and the answer\nlives in a stylesheet; whether the geometry a boundary would need exists at\nall is structural, and a component cannot conjure a shape from CSS. So the\npath is here for every donut and is `fill: none; stroke: none` until\nsomething asks for it.\n\n**Not a replacement for the dash.** The dash carries three jobs and only\none of them is the shape: a share growing is a `stroke-dasharray`\ninterpolating with no reflow, and `@starting-style` on that same property is\nthe whole reason the ring sweeps in rather than appearing. A `d` cannot do\neither. CSS interpolates one only between paths of identical command\nstructure, and the large-arc flag flips the moment a slice crosses a half\nturn, so a growing slice would pop partway through the transition it was\nmeant to smooth. Both animations stay on the dash and this path follows it.\n\nThe turn to twelve o'clock, or to nine for a half, is `origin` here rather\nthan the rotation the stylesheet gives the circles. That rule uses\n`transform-box: fill-box`, so it spins each element about its own bounding\nbox, which is the ring's center for a circle and is not for a sector."
             },
             {
               "kind": "method",
@@ -32899,7 +33258,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -33586,7 +33945,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -34234,7 +34593,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -34968,7 +35327,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -35488,7 +35847,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -36622,7 +36981,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -37324,7 +37683,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -37851,7 +38210,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -38578,7 +38937,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -39027,7 +39386,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -40767,7 +41126,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -41694,7 +42053,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -42403,6 +42762,26 @@
               "description": "The bubble's point, forwarded from `Tooltip`.",
               "name": "tooltip-arrow"
             },
+            {
+              "description": "The point while the bubble is below the reading, so on its top edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-bottom"
+            },
+            {
+              "description": "The point while the bubble is left of the reading, so on its right edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-left"
+            },
+            {
+              "description": "The point while the bubble is right of the reading, so on its left edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-right"
+            },
+            {
+              "description": "The point while the bubble is above the reading, so on its bottom edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-top"
+            },
+            {
+              "description": "The bubble while it is below the reading, beside `tooltip`.",
+              "name": "tooltip-bottom"
+            },
             {
               "description": "One column's name and value inside the tooltip.",
               "name": "tooltip-cell"
@@ -42411,10 +42790,22 @@
               "description": "The reading's own name inside the tooltip.",
               "name": "tooltip-label"
             },
+            {
+              "description": "The bubble while it is left of the reading, beside `tooltip`.",
+              "name": "tooltip-left"
+            },
             {
               "description": "A column's name inside the tooltip.",
               "name": "tooltip-name"
             },
+            {
+              "description": "The bubble while it is right of the reading, beside `tooltip`.",
+              "name": "tooltip-right"
+            },
+            {
+              "description": "The bubble while it is above the reading, beside `tooltip`.",
+              "name": "tooltip-top"
+            },
             {
               "description": "A column's value inside the tooltip.",
               "name": "tooltip-value"
@@ -42472,7 +42863,7 @@
                 "text": "number"
               },
               "default": "1",
-              "description": "How many rows the strip has.\n\n`1` is a status strip and `7` is a calendar week, which are the two readings\nthis covers. Cells fill down the column and then across, which is what a\ncalendar heatmap does: a week is a column, so the strip reads left to right\nin weeks and top to bottom in days.",
+              "description": "How many rows the strip has.\n\n`1` is a status strip and `7` is a calendar week, which are the two readings\nthis covers. Cells fill down the column and then across, which is what a\ncalendar heatmap does: a week is a column, so the strip reads along the\nline in weeks, from the right on a right-to-left page, and top to bottom in\ndays.",
               "attribute": "rows"
             },
             {
@@ -42903,7 +43294,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -43162,7 +43553,7 @@
                 "text": "number"
               },
               "default": "1",
-              "description": "How many rows the strip has.\n\n`1` is a status strip and `7` is a calendar week, which are the two readings\nthis covers. Cells fill down the column and then across, which is what a\ncalendar heatmap does: a week is a column, so the strip reads left to right\nin weeks and top to bottom in days.",
+              "description": "How many rows the strip has.\n\n`1` is a status strip and `7` is a calendar week, which are the two readings\nthis covers. Cells fill down the column and then across, which is what a\ncalendar heatmap does: a week is a column, so the strip reads along the\nline in weeks, from the right on a right-to-left page, and top to bottom in\ndays.",
               "fieldName": "rows"
             },
             {
@@ -44085,7 +44476,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -44373,6 +44764,16 @@
               "description": "Accessible name, for an icon that is not decorative.",
               "attribute": "label"
             },
+            {
+              "kind": "field",
+              "name": "mirror",
+              "type": {
+                "text": "boolean"
+              },
+              "default": "false",
+              "description": "Turns the glyph round on a right-to-left page, for an arrow meaning forward or back.\n\nAsked for rather than inferred from the name. The names are physical, and\n`arrow-left` drawn pointing right on an Arabic page would be the icon\ndisagreeing with its own name for every reader who meant the left.",
+              "attribute": "mirror"
+            },
             {
               "kind": "field",
               "name": "name",
@@ -44493,7 +44894,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -44718,6 +45119,15 @@
               "description": "Accessible name, for an icon that is not decorative.",
               "fieldName": "label"
             },
+            {
+              "name": "mirror",
+              "type": {
+                "text": "boolean"
+              },
+              "default": "false",
+              "description": "Turns the glyph round on a right-to-left page, for an arrow meaning forward or back.\n\nAsked for rather than inferred from the name. The names are physical, and\n`arrow-left` drawn pointing right on an Arabic page would be the icon\ndisagreeing with its own name for every reader who meant the left.",
+              "fieldName": "mirror"
+            },
             {
               "name": "name",
               "type": {
@@ -45160,7 +45570,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -47298,7 +47708,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -48056,7 +48466,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -48556,7 +48966,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -48935,6 +49345,26 @@
               "description": "The bubble's point, forwarded from `Tooltip`.",
               "name": "tooltip-arrow"
             },
+            {
+              "description": "The point while the bubble is below the reading, so on its top edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-bottom"
+            },
+            {
+              "description": "The point while the bubble is left of the reading, so on its right edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-left"
+            },
+            {
+              "description": "The point while the bubble is right of the reading, so on its left edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-right"
+            },
+            {
+              "description": "The point while the bubble is above the reading, so on its bottom edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-top"
+            },
+            {
+              "description": "The bubble while it is below the reading, beside `tooltip`.",
+              "name": "tooltip-bottom"
+            },
             {
               "description": "One column's name and value inside the tooltip.",
               "name": "tooltip-cell"
@@ -48943,10 +49373,22 @@
               "description": "The reading's own name inside the tooltip.",
               "name": "tooltip-label"
             },
+            {
+              "description": "The bubble while it is left of the reading, beside `tooltip`.",
+              "name": "tooltip-left"
+            },
             {
               "description": "A column's name inside the tooltip.",
               "name": "tooltip-name"
             },
+            {
+              "description": "The bubble while it is right of the reading, beside `tooltip`.",
+              "name": "tooltip-right"
+            },
+            {
+              "description": "The bubble while it is above the reading, beside `tooltip`.",
+              "name": "tooltip-top"
+            },
             {
               "description": "A column's value inside the tooltip.",
               "name": "tooltip-value"
@@ -49009,7 +49451,7 @@
               "type": {
                 "text": "'none' | 'ring' | 'square' | 'triangle' | undefined"
               },
-              "description": "The shape of the mark on each series' last reading.\n\nAbsent is a dot, and absent rather than defaulted because a reflecting\nproperty with a default writes its attribute onto a host the consumer\nrendered, which React reports as a hydration mismatch.\n\n**One per series, in a column at the right edge**, because every series here\nshares an x-axis and therefore ends at the same reading. That column is the\nuseful thing: it is where a reader looks for \"and what is it now\", which is\nthe question a sparkline's single mark answers too.\n\n`none` lets the lines stop on their own, which suits a chart read for its\nshape rather than for where it finished. The shapes come from\n`helix/styled/src/shared/chart/terminus.css`, the same sheet `Sparkline` takes, so\nthe set is identical wherever one appears.\n\n`ring` is the one that asks something of the drawing. It leaves a hole and a\nline ends at the reading, which is the hole's center, so the `<svg>` is\nmasked with one disc per series and every stroke stops at its own ring's\ninner edge. `mask-composite: intersect` is what makes N holes out of N\nlayers; a single `mask-image` would only ever cut the last one.",
+              "description": "The shape of the mark on each series' last reading.\n\nAbsent is a dot, and absent rather than defaulted because a reflecting\nproperty with a default writes its attribute onto a host the consumer\nrendered, which React reports as a hydration mismatch.\n\n**One per series, in a column at the far edge**, because every series here\nshares an x-axis and therefore ends at the same reading. That column is the\nuseful thing: it is where a reader looks for \"and what is it now\", which is\nthe question a sparkline's single mark answers too.\n\n`none` lets the lines stop on their own, which suits a chart read for its\nshape rather than for where it finished. The shapes come from\n`helix/styled/src/shared/chart/terminus.css`, the same sheet `Sparkline` takes, so\nthe set is identical wherever one appears.\n\n`ring` is the one that asks something of the drawing. It leaves a hole and a\nline ends at the reading, which is the hole's center, so the `<svg>` is\nmasked with one disc per series and every stroke stops at its own ring's\ninner edge. `mask-composite: intersect` is what makes N holes out of N\nlayers; a single `mask-image` would only ever cut the last one.",
               "attribute": "terminus",
               "reflects": true
             },
@@ -49485,7 +49927,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -49731,7 +50173,7 @@
               "type": {
                 "text": "'none' | 'ring' | 'square' | 'triangle' | undefined"
               },
-              "description": "The shape of the mark on each series' last reading.\n\nAbsent is a dot, and absent rather than defaulted because a reflecting\nproperty with a default writes its attribute onto a host the consumer\nrendered, which React reports as a hydration mismatch.\n\n**One per series, in a column at the right edge**, because every series here\nshares an x-axis and therefore ends at the same reading. That column is the\nuseful thing: it is where a reader looks for \"and what is it now\", which is\nthe question a sparkline's single mark answers too.\n\n`none` lets the lines stop on their own, which suits a chart read for its\nshape rather than for where it finished. The shapes come from\n`helix/styled/src/shared/chart/terminus.css`, the same sheet `Sparkline` takes, so\nthe set is identical wherever one appears.\n\n`ring` is the one that asks something of the drawing. It leaves a hole and a\nline ends at the reading, which is the hole's center, so the `<svg>` is\nmasked with one disc per series and every stroke stops at its own ring's\ninner edge. `mask-composite: intersect` is what makes N holes out of N\nlayers; a single `mask-image` would only ever cut the last one.",
+              "description": "The shape of the mark on each series' last reading.\n\nAbsent is a dot, and absent rather than defaulted because a reflecting\nproperty with a default writes its attribute onto a host the consumer\nrendered, which React reports as a hydration mismatch.\n\n**One per series, in a column at the far edge**, because every series here\nshares an x-axis and therefore ends at the same reading. That column is the\nuseful thing: it is where a reader looks for \"and what is it now\", which is\nthe question a sparkline's single mark answers too.\n\n`none` lets the lines stop on their own, which suits a chart read for its\nshape rather than for where it finished. The shapes come from\n`helix/styled/src/shared/chart/terminus.css`, the same sheet `Sparkline` takes, so\nthe set is identical wherever one appears.\n\n`ring` is the one that asks something of the drawing. It leaves a hole and a\nline ends at the reading, which is the hole's center, so the `<svg>` is\nmasked with one disc per series and every stroke stops at its own ring's\ninner edge. `mask-composite: intersect` is what makes N holes out of N\nlayers; a single `mask-image` would only ever cut the last one.",
               "fieldName": "terminus"
             },
             {
@@ -50045,7 +50487,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -50496,7 +50938,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -52025,7 +52467,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -52718,7 +53160,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -53547,7 +53989,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -54319,7 +54761,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -55299,7 +55741,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -56129,7 +56571,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -56697,7 +57139,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -57344,7 +57786,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -58050,7 +58492,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -58898,7 +59340,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -60341,7 +60783,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -60911,6 +61353,38 @@
               "description": "The point, resting against the trigger.",
               "name": "copy-tooltip-arrow"
             },
+            {
+              "description": "The point while the bubble is below the copy button, so on its top edge, beside `copy-tooltip-arrow`.",
+              "name": "copy-tooltip-arrow-bottom"
+            },
+            {
+              "description": "The point while the bubble is left of the copy button, so on its right edge, beside `copy-tooltip-arrow`.",
+              "name": "copy-tooltip-arrow-left"
+            },
+            {
+              "description": "The point while the bubble is right of the copy button, so on its left edge, beside `copy-tooltip-arrow`.",
+              "name": "copy-tooltip-arrow-right"
+            },
+            {
+              "description": "The point while the bubble is above the copy button, so on its bottom edge, beside `copy-tooltip-arrow`.",
+              "name": "copy-tooltip-arrow-top"
+            },
+            {
+              "description": "The bubble while it is below the copy button, beside `copy-tooltip`.",
+              "name": "copy-tooltip-bottom"
+            },
+            {
+              "description": "The bubble while it is left of the copy button, beside `copy-tooltip`.",
+              "name": "copy-tooltip-left"
+            },
+            {
+              "description": "The bubble while it is right of the copy button, beside `copy-tooltip`.",
+              "name": "copy-tooltip-right"
+            },
+            {
+              "description": "The bubble while it is above the copy button, beside `copy-tooltip`.",
+              "name": "copy-tooltip-top"
+            },
             {
               "description": "How many entries a branch holds.",
               "name": "count"
@@ -61414,7 +61888,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -61897,7 +62371,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -62492,7 +62966,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -63227,7 +63701,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -64172,7 +64646,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -64965,7 +65439,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -65471,7 +65945,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -65806,7 +66280,7 @@
       "declarations": [
         {
           "kind": "class",
-          "description": "**One `<input>` per cell, rather than one input drawn as cells.** A single\nfield autofills perfectly and needs a drawn caret nobody can style well;\nseparate fields give every box a real focus ring, a real caret and a real tab\nstop for free. Autofill's gap is closed below rather than lived with.\n\n**Anything longer than one character is treated as a paste, wherever it\nlands.** That covers the obvious case and two that are not: a browser\nautofilling an SMS code drops the whole thing into whichever box is focused,\nand a password manager does the same. So a six-character arrival in cell three\nfills from cell three rather than being truncated to its first character,\nwhich is what a per-cell `maxlength` would have done silently.\n\n**The code is one value to a form.** The cells are inside a shadow root and\nsubmit nothing on their own; the host is form-associated and reports the\njoined string, which is also what carries validity and answers a reset.\n\n`autocomplete=\"one-time-code\"` is on the first cell only. It is a hint about\nthe field as a whole, and repeating it on six boxes offers a browser six\nplaces to put one code.",
+          "description": "**One `<input>` per cell, rather than one input drawn as cells.** A single\nfield autofills perfectly and needs a drawn caret nobody can style well;\nseparate fields give every box a real focus ring, a real caret and a real tab\nstop for free. Autofill's gap is closed below rather than lived with.\n\n**Anything longer than one character is treated as a paste, wherever it\nlands.** That covers the obvious case and two that are not: a browser\nautofilling an SMS code drops the whole thing into whichever box is focused,\nand a password manager does the same. So a six-character arrival in cell three\nfills from cell three rather than being truncated to its first character,\nwhich is what a per-cell `maxlength` would have done silently.\n\n**The code is one value to a form.** The cells are inside a shadow root and\nsubmit nothing on their own; the host is form-associated and reports the\njoined string, which is also what carries validity and answers a reset.\n\n`autocomplete=\"one-time-code\"` is on the first cell only. It is a hint about\nthe field as a whole, and repeating it on six boxes offers a browser six\nplaces to put one code.\n\n**Left to right on every page.** A code is read the way a phone number is,\none character after another from the left, in Arabic and Hebrew as in\nEnglish, so the row carries `dir=\"ltr\"`: the first box is the leftmost, a\npaste fills rightward and the right arrow moves on. The label, the hint and\nthe feedback are words and keep the page's direction.",
           "name": "HelixHeadlessPinInput",
           "cssParts": [
             {
@@ -65818,7 +66292,7 @@
               "name": "feedback"
             },
             {
-              "description": "The row of boxes.",
+              "description": "The row of boxes, left to right on every page.",
               "name": "field"
             },
             {
@@ -66749,7 +67223,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -67610,7 +68084,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -68216,7 +68690,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -68790,7 +69264,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -69259,7 +69733,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -70277,7 +70751,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -70946,7 +71420,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -72231,7 +72705,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -72727,7 +73201,7 @@
       "declarations": [
         {
           "kind": "class",
-          "description": "The chart for a relation rather than a series. The table's first numeric\ncolumn is measured across and every other numeric column is measured up, one\npoint per row, so a table of countries with a column of income and a column\nof life expectancy draws one dot for each country. More numeric columns are\nmore series against the same across, each in its own color.\n\n**Both axes are measured, so both are labeled.** A line chart's readings are\nnamed along the bottom and its values are read from the tooltip; here the\nhorizontal position is itself a value, and a point with no scale under it says\nnothing about how far along it is. The ticks sit at round numbers and the\nrange is widened to them, so the plot's edges are numbers a reader can name.\n\n**What a reader without the picture is told is how strongly the two relate**,\nbecause that is the question a scatter is drawn to answer and the one a table\nof the same numbers makes hardest to see. The summary gives each series'\ncorrelation coefficient and how many points it rests on; the table has every\nnumber, and the drawing is `aria-hidden`.\n\n**The points are elements, not shapes in the drawing.** The drawing stretches\nto its box, and a circle in a stretched space is an ellipse, so each point is\na `<span>` placed by percentage over it, which is how `LineChart` draws its\ntermini and for the same reason.",
+          "description": "The chart for a relation rather than a series. The table's first numeric\ncolumn is measured across and every other numeric column is measured up, one\npoint per row, so a table of countries with a column of income and a column\nof life expectancy draws one dot for each country. More numeric columns are\nmore series against the same across, each in its own color.\n\n**Both axes are measured, so both are labeled.** A line chart's readings are\nnamed along the bottom and its values are read from the tooltip; here the\nhorizontal position is itself a value, and a point with no scale under it says\nnothing about how far along it is. The ticks sit at round numbers and the\nrange is widened to them, so the plot's edges are numbers a reader can name.\n\n**What a reader without the picture is told is how strongly the two relate**,\nbecause that is the question a scatter is drawn to answer and the one a table\nof the same numbers makes hardest to see. The summary gives each series'\ncorrelation coefficient and how many points it rests on; the table has every\nnumber, and the drawing is `aria-hidden`.\n\n**The points are elements, not shapes in the drawing.** The drawing stretches\nto its box, and a circle in a stretched space is an ellipse, so each point is\na `<span>` placed by percentage over it, which is how `LineChart` draws its\ntermini and for the same reason.\n\n**Left to right on every page, the one chart here that is.** A bar or a line\nchart's across is the order its readings are named in, which follows the\nline of text and turns with it. Here it is a measure, and a measure has a\ndirection of its own: more is to the right in every script, as a number line\nis. So the plot, its axes and the names at their ends sit in a `plane` part\nthat carries `dir=\"ltr\"`, which holds the y axis on the left and the points,\nrules, ticks, pointer and tooltip mark where a left-to-right page puts them,\nwith no stylesheet asked to know. The names carry `dir=\"auto\"`, so one written\nin Arabic or Hebrew still reads its own way; the summary is outside the plane\nand the bubble is told the page's direction, both being words.",
           "name": "HelixHeadlessScatterChart",
           "cssProperties": [
             {
@@ -72752,6 +73226,10 @@
               "description": "The invisible point the tooltip is anchored to.",
               "name": "mark"
             },
+            {
+              "description": "The plot, both axes and their names, left to right on every page. The grid they are laid out in.",
+              "name": "plane"
+            },
             {
               "description": "The drawing.",
               "name": "plot"
@@ -72784,6 +73262,26 @@
               "description": "The bubble's point, forwarded from `Tooltip`.",
               "name": "tooltip-arrow"
             },
+            {
+              "description": "The point while the bubble is below the reading, so on its top edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-bottom"
+            },
+            {
+              "description": "The point while the bubble is left of the reading, so on its right edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-left"
+            },
+            {
+              "description": "The point while the bubble is right of the reading, so on its left edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-right"
+            },
+            {
+              "description": "The point while the bubble is above the reading, so on its bottom edge, beside `tooltip-arrow`.",
+              "name": "tooltip-arrow-top"
+            },
+            {
+              "description": "The bubble while it is below the reading, beside `tooltip`.",
+              "name": "tooltip-bottom"
+            },
             {
               "description": "One column's name and value inside the tooltip.",
               "name": "tooltip-cell"
@@ -72792,10 +73290,22 @@
               "description": "The row's own name inside the tooltip.",
               "name": "tooltip-label"
             },
+            {
+              "description": "The bubble while it is left of the reading, beside `tooltip`.",
+              "name": "tooltip-left"
+            },
             {
               "description": "A column's name inside the tooltip.",
               "name": "tooltip-name"
             },
+            {
+              "description": "The bubble while it is right of the reading, beside `tooltip`.",
+              "name": "tooltip-right"
+            },
+            {
+              "description": "The bubble while it is above the reading, beside `tooltip`.",
+              "name": "tooltip-top"
+            },
             {
               "description": "A column's value inside the tooltip.",
               "name": "tooltip-value"
@@ -73192,7 +73702,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -73759,7 +74269,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -75102,7 +75612,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -75773,7 +76283,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -76243,7 +76753,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -77475,7 +77985,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -78208,7 +78718,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -78781,7 +79291,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -79440,7 +79950,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -80324,6 +80834,25 @@
               "readonly": true,
               "description": "The keys a separator answers.\n\nOne axis, the one the divider actually travels along, plus `Home` and `End`\nfor its bounds. The other two arrows are left alone and the page scrolls\nunder them as it would anywhere else.\n\n**Not all four, which is the tempting version.** Taking every arrow would\nmean `ArrowLeft` moving a horizontal divider up and down: not a reader\nbeing helped past a wrong guess, an arrow that does not go the way it\npoints. A reader who guesses wrong is answered better by the corner, where\nall four are live and each moves the divider it points along, which is the\nonly place in a splitter where a second axis means anything."
             },
+            {
+              "kind": "method",
+              "name": "travel",
+              "privacy": "private",
+              "return": {
+                "type": {
+                  "text": "-1 | 0 | 1"
+                }
+              },
+              "parameters": [
+                {
+                  "name": "key",
+                  "type": {
+                    "text": "string"
+                  }
+                }
+              ],
+              "description": "Which way a key moves the divider, as a sign on `position`, or `0` for a key\nthat is not along its axis.\n\n**The arrow says where the divider goes on screen**, the way the window\nsplitter pattern words it: left moves a side-by-side divider to the left.\n`position` is the start pane's share, and the start pane is on the right of\na right-to-left page, so there the left arrow grows it. It used to shrink\nit on every page, which on a right-to-left one sent the divider right."
+            },
             {
               "kind": "method",
               "name": "reset",
@@ -80562,7 +81091,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -81088,7 +81617,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -81533,7 +82062,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -82143,7 +82672,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -82756,7 +83285,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -83803,7 +84332,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -84248,6 +84777,12 @@
           "kind": "class",
           "description": "Carries its own role, because a grid of divs is not a table to anything\nreading it. `header` makes the cell a `columnheader`, which is what lets a\nscreen reader announce \"Attribute, appearance\" rather than just the value.\n\nThe role goes on the host rather than inside the shadow root, because the\ncell has to be a child of its row in the flat tree for the table's structure\nto hold together.",
           "name": "HelixHeadlessTableCell",
+          "cssProperties": [
+            {
+              "description": "The edge the column this cell sits under is set to, `start`, `center` or `end`, and unset under a column that sets none. Written by the table around it, since the cell is the consumer's markup and nothing is written onto it; read by the sheet that draws the cell, where `justify-content` and `text-align` both take it as it comes and both follow the writing direction.",
+              "name": "--table-cell-align"
+            }
+          ],
           "slots": [
             {
               "description": "The cell's content.",
@@ -84376,7 +84911,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -84684,7 +85219,7 @@
         },
         {
           "kind": "class",
-          "description": "**The column is the source of both its heading and its track.** Add one and\nthe table has another track; remove one and it does not. Declaring the widths\nseparately, as a CSS track list on the table, leaves nothing tying the two\ntogether: a table with three headings and a two-track list renders without\ncomplaint, and a track list that is accidentally the headings is invalid CSS\nthe grid silently ignores, collapsing the whole table into one column.\n\n**It is also where every per-column fact goes**, which is the reason\n`helix-data-table` needs no column config object. `field` names the key the\ncell reads, `sortable` says a reader may reorder by it, `align` says which\nedge the values line up against, and `cell` takes a function for anything the\nfield alone cannot express. The schema is markup and the data is a property:\na schema is small, declarative, and something a reader should be able to see\nin the HTML, and data is large and windowed.\n\n**The same element serves both tables**, the markup one and the data one. It\nis identified by a flag rather than by `instanceof`, so a `helix-table-column`\nworks inside either without the two roots sharing anything else.\n\n**It generates no box of its own**, the same as a row: the heading is a grid\nitem of the table, so it lines up with the cells under it rather than sitting\nin a wrapper that has to be kept the same width.\n\n**A heading that does something holds a control that does it.** The host is\nthe column header, and a header is not focusable, so a sortable heading is a\n`<button>` inside it: the pattern for a sortable table, and the only way a\nreader on a keyboard reaches the sort at all. A column a reader may resize\ncarries a `separator` on its end edge, moved by dragging it or with the arrow\nkeys, and a column in a table that lets a reader reorder its columns carries\na grip the table moves.",
+          "description": "**The column is the source of both its heading and its track.** Add one and\nthe table has another track; remove one and it does not. Declaring the widths\nseparately, as a CSS track list on the table, leaves nothing tying the two\ntogether: a table with three headings and a two-track list renders without\ncomplaint, and a track list that is accidentally the headings is invalid CSS\nthe grid silently ignores, collapsing the whole table into one column.\n\n**It is also where every per-column fact goes**, which is the reason\n`helix-data-table` needs no column config object. `field` names the key the\ncell reads, `sortable` says a reader may reorder by it, `align` says which\nedge the values line up against, and `cell` takes a function for anything the\nfield alone cannot express. The schema is markup and the data is a property:\na schema is small, declarative, and something a reader should be able to see\nin the HTML, and data is large and windowed.\n\n**The same element serves both tables**, the markup one and the data one. It\nis identified by a flag rather than by `instanceof`, so a `helix-table-column`\nworks inside either without the two roots sharing anything else.\n\n**It generates no box of its own**, the same as a row: the heading is a grid\nitem of the table, so it lines up with the cells under it rather than sitting\nin a wrapper that has to be kept the same width.\n\n**`hidden` takes the column out of the grid**, its heading and its track\nboth, so the columns after it keep their own widths and every row keeps its\ncells under their headings. `hidden=\"until-found\"` keeps its track, the\nheading being left drawn, as every Helix element leaves one with that value.\n`helix-data-table` draws the cells, so it leaves out the ones under a hidden\ncolumn. A `helix-table`'s cells are markup the consumer wrote, so hide them\nwith the column, as removing a column means removing its cells: a row with a\ncell more than the table has tracks wraps into the next.\n\n**A heading that does something holds a control that does it.** The host is\nthe column header, and a header is not focusable, so a sortable heading is a\n`<button>` inside it: the pattern for a sortable table, and the only way a\nreader on a keyboard reaches the sort at all. A column a reader may resize\ncarries a `separator` on its end edge, moved by dragging it or with the arrow\nkeys, and a column in a table that lets a reader reorder its columns carries\na grip the table moves.",
           "name": "HelixHeadlessTableColumn",
           "cssParts": [
             {
@@ -84721,7 +85256,7 @@
               "type": {
                 "text": "'center' | 'end' | 'start' | undefined"
               },
-              "description": "Which edge the values line up against.\n\n`end` is what a column of numbers wants, so digits of the same magnitude sit\nabove each other. Left unset the cell inherits the table's own alignment,\nwhich is `start` for text in either writing direction. `helix-data-table`\naligns the cells it draws; a `helix-table`'s cells are markup the consumer\nwrote, and align as the consumer styles them. The heading is this element's\nown, and `@fusion.dev/helix` sets it to the same edge in either table.\n\n**Reflected, because the heading is drawn by a stylesheet and has no other\nway to learn it.** Set as a property, which is how React sets it, an\nunreflected `align` left no attribute to select on, so the heading stayed at\nthe start over a column of digits set to the end. It is the author's\ndeclaration and nothing moves it afterwards, so an attribute cannot come to\ndisagree with the element, which is the reason `open` records a custom\nstate instead. No default, so an unset column writes nothing onto its host.",
+              "description": "Which edge the values line up against.\n\n`end` is what a column of numbers wants, so digits of the same magnitude sit\nabove each other. Left unset the cell inherits the table's own alignment,\nwhich is `start` for text in either writing direction. `helix-data-table`\naligns the cells it draws. A `helix-table`'s cells are markup the consumer\nwrote, so the table writes nothing onto them: it publishes the edge to each\ncell under this column as `--table-cell-align`, which `@fusion.dev/helix`\nreads. The heading is this element's own, and `@fusion.dev/helix` sets it to\nthe same edge in either table. `end` is the inline end, so in a\nright-to-left table it is the left.\n\n**Reflected, because the heading is drawn by a stylesheet and has no other\nway to learn it.** Set as a property, which is how React sets it, an\nunreflected `align` left no attribute to select on, so the heading stayed at\nthe start over a column of digits set to the end. It is the author's\ndeclaration and nothing moves it afterwards, so an attribute cannot come to\ndisagree with the element, which is the reason `open` records a custom\nstate instead. No default, so an unset column writes nothing onto its host.",
               "attribute": "align",
               "reflects": true
             },
@@ -84770,8 +85305,9 @@
                 "text": "boolean"
               },
               "default": "false",
-              "description": "Whether a reader may reorder the set by this column.\n\nIgnored without a `field`, and silently: a column with nothing to sort by\nhas no order to offer, and refusing loudly would mean an actions column\ncarrying `sortable` breaks a table rather than doing nothing.",
-              "attribute": "sortable"
+              "description": "Whether a reader may reorder the set by this column.\n\nIgnored without a `field`, and silently: a column with nothing to sort by\nhas no order to offer, and refusing loudly would mean an actions column\ncarrying `sortable` breaks a table rather than doing nothing.\n\n**Reflected, because a stylesheet selects on it.** `@fusion.dev/helix`\ngives a data table's sortable headings a pointer through\n`::slotted(helix-table-column[sortable])`, and set as a property, which is\nhow React sets it, an unreflected `sortable` left no attribute to match.\n`false` reflects as no attribute at all, so an unset column writes nothing\nonto its host.",
+              "attribute": "sortable",
+              "reflects": true
             },
             {
               "kind": "field",
@@ -84780,7 +85316,7 @@
                 "text": "string"
               },
               "default": "''",
-              "description": "The track this column takes, as CSS.\n\nA single track value, so anything valid in `grid-template-columns` works:\n`16rem`, `auto`, `minmax(0, 2fr)`. Left unset the column shares what is\nleft over equally, as `minmax(max-content, 1fr)`.\n\n**The floor is `max-content`, never `0` and never `min-content`.** A zero\nfloor lets a track shrink under its own content: at 375 the columns squeeze\nto 75px and `eu-central-1` wraps across lines, which reads as corrupt data\nrather than as a table that ran out of room. A `min-content` floor still\nbreaks it, a hyphen being a wrap opportunity.\n\nThe overflow has somewhere to go: `@fusion.dev/helix` scrolls the table's\nown box, so a long value pushes past the container and the table scrolls\ninstead of the value being destroyed. Measured at 375: tracks of 192, 89 and\n102 against a 343 box, so the table scrolls 384 and the page does not move.\n\nThe cost is that a column of prose asks for its longest paragraph on one\nline, so a column holding sentences rather than values wants\n`width=\"minmax(0, 1fr)\"` to opt back into wrapping. A headless consumer owns\nthe overflow either way, that being paint, and gets the honest track list.",
+              "description": "The track this column takes, as CSS.\n\nA single track value, so anything valid in `grid-template-columns` works:\n`16rem`, `auto`, `minmax(0, 2fr)`. Left unset the column takes the track\nin its place in a `helix-table`'s `columns`, where that table sets one, and\notherwise shares what is left over equally, as `minmax(max-content, 1fr)`.\nSet here, it wins over that list, which is what lets a reader resize a\ncolumn in a table that set one.\n\n**The floor is `max-content`, never `0` and never `min-content`.** A zero\nfloor lets a track shrink under its own content: at 375 the columns squeeze\nto 75px and `eu-central-1` wraps across lines, which reads as corrupt data\nrather than as a table that ran out of room. A `min-content` floor still\nbreaks it, a hyphen being a wrap opportunity.\n\nThe overflow has somewhere to go: `@fusion.dev/helix` scrolls the table's\nown box, so a long value pushes past the container and the table scrolls\ninstead of the value being destroyed. Measured at 375: tracks of 192, 89 and\n102 against a 343 box, so the table scrolls 384 and the page does not move.\n\nThe cost is that a column of prose asks for its longest paragraph on one\nline, so a column holding sentences rather than values wants\n`width=\"minmax(0, 1fr)\"` to opt back into wrapping. A headless consumer owns\nthe overflow either way, that being paint, and gets the honest track list.",
               "attribute": "width"
             },
             {
@@ -84811,6 +85347,15 @@
               "privacy": "private",
               "description": "Watches the heading's width, which changes with the table as well as with a drag."
             },
+            {
+              "kind": "field",
+              "name": "hiding",
+              "type": {
+                "text": "MutationObserver | undefined"
+              },
+              "privacy": "private",
+              "description": "Watches `hidden`, which is the platform's attribute and no property of this element's."
+            },
             {
               "kind": "field",
               "name": "NARROWEST",
@@ -84913,6 +85458,17 @@
                 }
               }
             },
+            {
+              "kind": "method",
+              "name": "tellHolder",
+              "privacy": "private",
+              "return": {
+                "type": {
+                  "text": "void"
+                }
+              },
+              "description": "Asks whatever holds this column to render, since it reads the column when it does."
+            },
             {
               "kind": "method",
               "name": "resizeTo",
@@ -84939,7 +85495,7 @@
                 "text": "number"
               },
               "privacy": "private",
-              "description": "Which way a drag or an arrow key grows the column, in this writing direction.",
+              "description": "Which way a drag grows the column, in this writing direction.",
               "readonly": true
             },
             {
@@ -85077,7 +85633,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -85308,7 +85864,7 @@
               "type": {
                 "text": "'center' | 'end' | 'start' | undefined"
               },
-              "description": "Which edge the values line up against.\n\n`end` is what a column of numbers wants, so digits of the same magnitude sit\nabove each other. Left unset the cell inherits the table's own alignment,\nwhich is `start` for text in either writing direction. `helix-data-table`\naligns the cells it draws; a `helix-table`'s cells are markup the consumer\nwrote, and align as the consumer styles them. The heading is this element's\nown, and `@fusion.dev/helix` sets it to the same edge in either table.\n\n**Reflected, because the heading is drawn by a stylesheet and has no other\nway to learn it.** Set as a property, which is how React sets it, an\nunreflected `align` left no attribute to select on, so the heading stayed at\nthe start over a column of digits set to the end. It is the author's\ndeclaration and nothing moves it afterwards, so an attribute cannot come to\ndisagree with the element, which is the reason `open` records a custom\nstate instead. No default, so an unset column writes nothing onto its host.",
+              "description": "Which edge the values line up against.\n\n`end` is what a column of numbers wants, so digits of the same magnitude sit\nabove each other. Left unset the cell inherits the table's own alignment,\nwhich is `start` for text in either writing direction. `helix-data-table`\naligns the cells it draws. A `helix-table`'s cells are markup the consumer\nwrote, so the table writes nothing onto them: it publishes the edge to each\ncell under this column as `--table-cell-align`, which `@fusion.dev/helix`\nreads. The heading is this element's own, and `@fusion.dev/helix` sets it to\nthe same edge in either table. `end` is the inline end, so in a\nright-to-left table it is the left.\n\n**Reflected, because the heading is drawn by a stylesheet and has no other\nway to learn it.** Set as a property, which is how React sets it, an\nunreflected `align` left no attribute to select on, so the heading stayed at\nthe start over a column of digits set to the end. It is the author's\ndeclaration and nothing moves it afterwards, so an attribute cannot come to\ndisagree with the element, which is the reason `open` records a custom\nstate instead. No default, so an unset column writes nothing onto its host.",
               "fieldName": "align"
             },
             {
@@ -85335,7 +85891,7 @@
                 "text": "boolean"
               },
               "default": "false",
-              "description": "Whether a reader may reorder the set by this column.\n\nIgnored without a `field`, and silently: a column with nothing to sort by\nhas no order to offer, and refusing loudly would mean an actions column\ncarrying `sortable` breaks a table rather than doing nothing.",
+              "description": "Whether a reader may reorder the set by this column.\n\nIgnored without a `field`, and silently: a column with nothing to sort by\nhas no order to offer, and refusing loudly would mean an actions column\ncarrying `sortable` breaks a table rather than doing nothing.\n\n**Reflected, because a stylesheet selects on it.** `@fusion.dev/helix`\ngives a data table's sortable headings a pointer through\n`::slotted(helix-table-column[sortable])`, and set as a property, which is\nhow React sets it, an unreflected `sortable` left no attribute to match.\n`false` reflects as no attribute at all, so an unset column writes nothing\nonto its host.",
               "fieldName": "sortable"
             },
             {
@@ -85344,7 +85900,7 @@
                 "text": "string"
               },
               "default": "''",
-              "description": "The track this column takes, as CSS.\n\nA single track value, so anything valid in `grid-template-columns` works:\n`16rem`, `auto`, `minmax(0, 2fr)`. Left unset the column shares what is\nleft over equally, as `minmax(max-content, 1fr)`.\n\n**The floor is `max-content`, never `0` and never `min-content`.** A zero\nfloor lets a track shrink under its own content: at 375 the columns squeeze\nto 75px and `eu-central-1` wraps across lines, which reads as corrupt data\nrather than as a table that ran out of room. A `min-content` floor still\nbreaks it, a hyphen being a wrap opportunity.\n\nThe overflow has somewhere to go: `@fusion.dev/helix` scrolls the table's\nown box, so a long value pushes past the container and the table scrolls\ninstead of the value being destroyed. Measured at 375: tracks of 192, 89 and\n102 against a 343 box, so the table scrolls 384 and the page does not move.\n\nThe cost is that a column of prose asks for its longest paragraph on one\nline, so a column holding sentences rather than values wants\n`width=\"minmax(0, 1fr)\"` to opt back into wrapping. A headless consumer owns\nthe overflow either way, that being paint, and gets the honest track list.",
+              "description": "The track this column takes, as CSS.\n\nA single track value, so anything valid in `grid-template-columns` works:\n`16rem`, `auto`, `minmax(0, 2fr)`. Left unset the column takes the track\nin its place in a `helix-table`'s `columns`, where that table sets one, and\notherwise shares what is left over equally, as `minmax(max-content, 1fr)`.\nSet here, it wins over that list, which is what lets a reader resize a\ncolumn in a table that set one.\n\n**The floor is `max-content`, never `0` and never `min-content`.** A zero\nfloor lets a track shrink under its own content: at 375 the columns squeeze\nto 75px and `eu-central-1` wraps across lines, which reads as corrupt data\nrather than as a table that ran out of room. A `min-content` floor still\nbreaks it, a hyphen being a wrap opportunity.\n\nThe overflow has somewhere to go: `@fusion.dev/helix` scrolls the table's\nown box, so a long value pushes past the container and the table scrolls\ninstead of the value being destroyed. Measured at 375: tracks of 192, 89 and\n102 against a 343 box, so the table scrolls 384 and the page does not move.\n\nThe cost is that a column of prose asks for its longest paragraph on one\nline, so a column holding sentences rather than values wants\n`width=\"minmax(0, 1fr)\"` to opt back into wrapping. A headless consumer owns\nthe overflow either way, that being paint, and gets the honest track list.",
               "fieldName": "width"
             }
           ],
@@ -85505,7 +86061,7 @@
                 "text": "string"
               },
               "default": "''",
-              "description": "Column track list, as CSS. Deprecated: declare columns instead.\n\nKept for a table that has no `helix-table-column` children, and ignored the\nmoment it has one. A raw track list is the right shape for the value and\nthe wrong place for it: the headings and the widths were two halves of one\nfact held in two places, and nothing checked that they agreed. Worse, the\nvalue is CSS, so handing it headings by mistake produced an invalid track\nlist, which a grid ignores in silence: the table collapsed to one column\nand said nothing. Columns carry both halves now.",
+              "description": "The column widths as one track list, written once on the table.\n\nWith `helix-table-column` children, each track sizes the column in its\nplace, a hidden column counted, so hiding one takes its track with it and\nthe rest keep theirs. The columns decide how many tracks there are and this\nlist only how wide: a track past the last column is dropped, and a column\npast the last track shares what is left. With no columns declared it is\nthe whole track list. Tracks count as the grid counts them, so\n`minmax(0, 1fr)` is one and `repeat(2, 8rem)` is two, and one the grid\nwould refuse leaves its column at the default rather than taking the table\ndown to one column.\n\n**A column's own `width` wins over its track here.** The width is written\non the column it sizes, which is the more particular statement, and it is\nwhere a resize lands: a reader dragging a column's edge writes `width`, so\nwere this list to win, the drag would move the separator and leave the\ntrack where the list put it. This is the table's default for its columns,\nand each column may say otherwise.",
               "attribute": "columns"
             },
             {
@@ -85534,7 +86090,7 @@
                 "text": "string"
               },
               "privacy": "private",
-              "description": "The track list, from the columns if there are any.\n\nRead from the DOM rather than kept, so adding or removing a column needs\nnothing else to be told. Falls back to the `columns` string for a table\nwritten before columns existed, and to a single track for one that declares\nneither.",
+              "description": "The track list, from the columns if there are any.\n\nRead from the DOM rather than kept, so adding or removing a column needs\nnothing else to be told. A hidden column takes no track: left in, the\nheader row had a heading fewer than the grid had tracks, the first cell\nfilled the gap, and every row after it began a column late. Falls back to\nthe `columns` string for a table that declares no columns, and to a single\ntrack for one that declares neither.",
               "readonly": true
             },
             {
@@ -85543,6 +86099,16 @@
               "privacy": "private",
               "readonly": true
             },
+            {
+              "kind": "field",
+              "name": "watcher",
+              "type": {
+                "text": "MutationObserver | undefined"
+              },
+              "privacy": "private",
+              "default": "undefined",
+              "description": "Watches the light DOM for rows and columns arriving, leaving and moving.\n\n**A manually assigned slot does not announce them**, the same gap `Chat`\nand `Carousel` watch for. Under `slotAssignment: 'manual'` only `assign`\nchanges an assignment, so a row appended after the first render reached no\nslot and drew nowhere, and a column moved to another place lost its\nassignment on the way out and never got it back: measured in Chromium, both\n`assignedSlot` null a render later. A table whose definition loaded before\nits markup was parsed drew no columns at all, having rendered before any\narrived."
+            },
             {
               "kind": "field",
               "name": "shadowRootOptions",
@@ -85577,6 +86143,37 @@
                 }
               }
             },
+            {
+              "kind": "field",
+              "name": "alignment",
+              "type": {
+                "text": "CSSStyleSheet | undefined"
+              },
+              "privacy": "private",
+              "default": "undefined",
+              "description": "The rules handing each column's `align` to the cells under it. One sheet for\nevery row this table holds, rewritten in place, so a column changing its\nedge reaches every row in one write."
+            },
+            {
+              "kind": "field",
+              "name": "aligned",
+              "type": {
+                "text": "string"
+              },
+              "privacy": "private",
+              "default": "''",
+              "description": "What `alignment` was last written with, so a render that changed no edge writes nothing."
+            },
+            {
+              "kind": "method",
+              "name": "alignCells",
+              "privacy": "private",
+              "return": {
+                "type": {
+                  "text": "void"
+                }
+              },
+              "description": "Tells every row which edge each of its cells lines up against.\n\n**A rule in each row's shadow root, keyed on position and written by this\ntable.** The cells are the consumer's markup, so nothing may be written onto\nthem, and each of the other ways fails somewhere a table goes:\n\n- This element's own sheet cannot reach them. `::slotted()` matches what is\n  assigned here, which is the rows, and never a row's children.\n- A property per column published here, `--table-align-2: end`, still needs\n  a rule that knows which column a cell is in, and a fixed ladder of\n  `:nth-child` rules in a static sheet is a ceiling on the column count.\n- A custom state set on each cell by script is right only until a cell\n  moves, and a cell moving inside a row changes no child list this element\n  watches.\n\n`:nth-child` is matched by the browser against wherever a cell is now, so a\ncell that moves takes its new column's edge with no script at all; a column\nthat moves rewrites the rules on the render the watcher asks for; and a\ncolumn hidden along with its cells keeps its place in both counts, since\n`:nth-child` counts a hidden sibling the way `tableColumns` counts a hidden\ncolumn. The row is where they go because only the shadow root holding a\nslot can style what is assigned to it.\n\n**A custom property and nothing else**, which is how this tier publishes a\nfact: the edge is the author's declaration, and what lining up means, a\n`justify-content` or a `text-align`, belongs to whoever draws the cell.\n`initial` on every cell first, so a table nested in an aligned cell does not\ntake the outer column's edge for columns of its own that set none."
+            },
             {
               "kind": "field",
               "name": "emptyText",
@@ -85817,7 +86414,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -86039,7 +86636,7 @@
                 "text": "string"
               },
               "default": "''",
-              "description": "Column track list, as CSS. Deprecated: declare columns instead.\n\nKept for a table that has no `helix-table-column` children, and ignored the\nmoment it has one. A raw track list is the right shape for the value and\nthe wrong place for it: the headings and the widths were two halves of one\nfact held in two places, and nothing checked that they agreed. Worse, the\nvalue is CSS, so handing it headings by mistake produced an invalid track\nlist, which a grid ignores in silence: the table collapsed to one column\nand said nothing. Columns carry both halves now.",
+              "description": "The column widths as one track list, written once on the table.\n\nWith `helix-table-column` children, each track sizes the column in its\nplace, a hidden column counted, so hiding one takes its track with it and\nthe rest keep theirs. The columns decide how many tracks there are and this\nlist only how wide: a track past the last column is dropped, and a column\npast the last track shares what is left. With no columns declared it is\nthe whole track list. Tracks count as the grid counts them, so\n`minmax(0, 1fr)` is one and `repeat(2, 8rem)` is two, and one the grid\nwould refuse leaves its column at the default rather than taking the table\ndown to one column.\n\n**A column's own `width` wins over its track here.** The width is written\non the column it sizes, which is the more particular statement, and it is\nwhere a resize lands: a reader dragging a column's edge writes `width`, so\nwere this list to win, the drag would move the separator and leave the\ntrack where the list put it. This is the table's default for its columns,\nand each column may say otherwise.",
               "fieldName": "columns"
             },
             {
@@ -86172,6 +86769,16 @@
             }
           ],
           "members": [
+            {
+              "kind": "field",
+              "name": "columnRules",
+              "type": {
+                "text": "CSSStyleSheet | undefined"
+              },
+              "privacy": "private",
+              "default": "undefined",
+              "description": "The table's rules this row holds, so the next ones replace them rather than pile up."
+            },
             {
               "kind": "field",
               "name": "tier",
@@ -86282,7 +86889,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -86740,7 +87347,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -87372,7 +87979,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -87920,7 +88527,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -88448,7 +89055,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -89741,7 +90348,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -90841,13 +91448,36 @@
               "name": "panel",
               "privacy": "protected",
               "readonly": true,
-              "default": "new PopoverController(this, { anchor: () => this.partNamed('field'), /* The surface *and* the trigger. * * A control is not always one element: the button that opens this panel is * part of it, and a press on it is not a press outside. Scoped to the * surface alone, `pointerdown` on the trigger closes the panel and the * trigger's own `toggleOpen` reopens it on the `click` that follows, so the * panel can never be closed by the button that opened it. * `DismissController` documents this failure and takes a list for it. */ boundary: (): readonly (Element | null | undefined)[] => [ this.partNamed('surface'), this.partNamed('trigger'), ], /* Wider than the four every other panel here takes. Those sit against a control the same width as themselves and read as an extension of it; this one is a picker of its own size hanging off a field, so the gap is what says the two are separate things rather than one control that grew. */ offset: () => 8, onDismiss: () => this.requestOpen(false), open: () => this.open, placement: 'bottom-start', surface: () => this.partNamed('surface'), })",
+              "default": "new PopoverController(this, { anchor: () => this.partNamed('field'), /* The surface *and* the trigger. * * A control is not always one element: the button that opens this panel is * part of it, and a press on it is not a press outside. Scoped to the * surface alone, `pointerdown` on the trigger closes the panel and the * trigger's own `toggleOpen` reopens it on the `click` that follows, so the * panel can never be closed by the button that opened it. * `DismissController` documents this failure and takes a list for it. */ boundary: (): readonly (Element | null | undefined)[] => [ this.partNamed('surface'), this.partNamed('trigger'), ], /* Wider than the four every other panel here takes. Those sit against a control the same width as themselves and read as an extension of it; this one is a picker of its own size hanging off a field, so the gap is what says the two are separate things rather than one control that grew. */ offset: () => 8, onDismiss: () => this.requestOpen(false), onPlace: (_placement, room) => this.placed(room), open: () => this.open, placement: 'bottom-start', surface: () => this.partNamed('surface'), })",
               "description": "The overlay, placed against the field rather than the trigger.\n\nAgainst the field because that is the box a reader sees as the control: a\ncalendar hung off a small button at one end of a wide field reads as\nbelonging to the button. `Select` anchors to its trigger for the opposite\nreason, its trigger being the whole control.",
               "inheritedFrom": {
                 "name": "MixinPickerField",
                 "module": "src/mixins/mixin-picker-field/mixin-picker-field.ts"
               }
             },
+            {
+              "kind": "method",
+              "name": "placed",
+              "privacy": "protected",
+              "return": {
+                "type": {
+                  "text": "void"
+                }
+              },
+              "parameters": [
+                {
+                  "name": "_room",
+                  "type": {
+                    "text": "number | undefined"
+                  }
+                }
+              ],
+              "description": "Told the panel has been placed, and how wide the window leaves it room to\nbe, in pixels.\n\nNothing, unless a family fits what its panel holds to the room it has.",
+              "inheritedFrom": {
+                "name": "MixinPickerField",
+                "module": "src/mixins/mixin-picker-field/mixin-picker-field.ts"
+              }
+            },
             {
               "kind": "field",
               "name": "anchor",
@@ -91890,7 +92520,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -93597,7 +94227,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -94300,7 +94930,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -94723,7 +95353,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -95415,7 +96045,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -95897,7 +96527,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -96943,7 +97573,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -97607,7 +98237,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -98173,7 +98803,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -98495,9 +99125,41 @@
               "description": "The point, resting against the trigger.",
               "name": "arrow"
             },
+            {
+              "description": "The point while the bubble is below its trigger, so on the bubble's top edge, beside `arrow`.",
+              "name": "arrow-bottom"
+            },
+            {
+              "description": "The point while the bubble is left of its trigger, so on the bubble's right edge, beside `arrow`.",
+              "name": "arrow-left"
+            },
+            {
+              "description": "The point while the bubble is right of its trigger, so on the bubble's left edge, beside `arrow`.",
+              "name": "arrow-right"
+            },
+            {
+              "description": "The point while the bubble is above its trigger, so on the bubble's bottom edge, beside `arrow`.",
+              "name": "arrow-top"
+            },
             {
               "description": "The bubble.",
               "name": "tooltip"
+            },
+            {
+              "description": "The bubble while it is below its trigger, beside `tooltip`.",
+              "name": "tooltip-bottom"
+            },
+            {
+              "description": "The bubble while it is left of its trigger, beside `tooltip`.",
+              "name": "tooltip-left"
+            },
+            {
+              "description": "The bubble while it is right of its trigger, beside `tooltip`.",
+              "name": "tooltip-right"
+            },
+            {
+              "description": "The bubble while it is above its trigger, beside `tooltip`.",
+              "name": "tooltip-top"
             }
           ],
           "slots": [
@@ -98603,6 +99265,15 @@
               "description": "The id of a box to keep the bubble inside, instead of the viewport.\n\nResolved the way `for` is, so an id in the same tree, which is what lets a\ncomponent name a box in its own shadow root. A chart does: its bubble is\ncentered on a reading, and the first and last readings sit at the ends of\nthe plot, so half the box hangs past the drawing and, measured on a chart\nin a 320px card, three pixels past the card itself.\n\nClamping is on the cross axis only, which is what makes this safe to point\nat a box the bubble genuinely does not fit inside. See `within` on\n`FloatingController` for what happens if it is applied to the rest.",
               "attribute": "within"
             },
+            {
+              "kind": "field",
+              "name": "side",
+              "type": {
+                "text": "Side | undefined"
+              },
+              "privacy": "private",
+              "description": "The side of its trigger the bubble last landed on, which is not always\n`placement`: a bubble with no room there is moved.\n\nUndefined until the first placement, which the server never makes, so a\npart name it adds is never markup the server did not send."
+            },
             {
               "kind": "field",
               "name": "timer",
@@ -98657,7 +99328,7 @@
               "name": "bubbleLayer",
               "privacy": "private",
               "readonly": true,
-              "default": "new PopoverController(this, { anchor: () => this.trigger, arrow: () => this.partNamed('arrow'), /* Annotated, because without it the array's element type is inferred from `this`, whose own type depends on this controller, and TypeScript stops at the cycle. */ boundary: (): readonly (Element | null)[] => [this, this.named], /* The declared gap plus the arrow's own reach, so the gap is to the point. * * A square of side `n` rotated forty-five degrees about its own center * reaches `n / sqrt(2)` from that center, and the arrow is placed with its * center on the bubble's edge, so that is exactly how far past the edge the * tip goes. Read from the element rather than assumed, because * `--tooltip-arrow-size` is a hook and a theme may have moved it. * * `offsetWidth` and not `getBoundingClientRect`, which reports the rotated * box and would count the reach twice. Zero while the bubble is closed and * unrendered, which is right: there is nothing to place then either. */ offset: () => { const point = this.partNamed('arrow') return this.offset + (point?.offsetWidth ?? 0) / Math.SQRT2 }, /* Escape closes the bubble from wherever the reader is, and this is the one listener that hears it. On the document, through the controller, not on the host: a tooltip opened by hover has focus somewhere else entirely, so a press never reaches this element and Escape would work for everything except the case the pattern is named for, a reader escaping a tooltip obscuring what is under it. The controller listens only while the bubble is up, and it is the same arbitration every overlay goes through, so a press with a deeper surface open closes that one and leaves this alone rather than taking both: a second listener of this element's own answered every press for the element's whole connected life and skipped the vote. Holds are cleared as well as hiding, for the reason `handleActivate` gives: left set, the pointer is still recorded as being over the trigger, and the bubble comes back on a `release` it should have been dropped from. */ onDismiss: () => { this.holds.clear() this.hide() }, placement: () => this.placement, surface: () => this.partNamed('tooltip'), within: () => this.enclosure, })"
+              "default": "new PopoverController(this, { anchor: () => this.trigger, arrow: () => this.partNamed('arrow'), /* Annotated, because without it the array's element type is inferred from `this`, whose own type depends on this controller, and TypeScript stops at the cycle. */ boundary: (): readonly (Element | null)[] => [this, this.named], /* The declared gap plus the arrow's own reach, so the gap is to the point. * * A square of side `n` rotated forty-five degrees about its own center * reaches `n / sqrt(2)` from that center, and the arrow is placed with its * center on the bubble's edge, so that is exactly how far past the edge the * tip goes. Read from the element rather than assumed, because * `--tooltip-arrow-size` is a hook and a theme may have moved it. * * `offsetWidth` and not `getBoundingClientRect`, which reports the rotated * box and would count the reach twice. Zero while the bubble is closed and * unrendered, which is right: there is nothing to place then either. */ offset: () => { const point = this.partNamed('arrow') return this.offset + (point?.offsetWidth ?? 0) / Math.SQRT2 }, /* Escape closes the bubble from wherever the reader is, and this is the one listener that hears it. On the document, through the controller, not on the host: a tooltip opened by hover has focus somewhere else entirely, so a press never reaches this element and Escape would work for everything except the case the pattern is named for, a reader escaping a tooltip obscuring what is under it. The controller listens only while the bubble is up, and it is the same arbitration every overlay goes through, so a press with a deeper surface open closes that one and leaves this alone rather than taking both: a second listener of this element's own answered every press for the element's whole connected life and skipped the vote. Holds are cleared as well as hiding, for the reason `handleActivate` gives: left set, the pointer is still recorded as being over the trigger, and the bubble comes back on a `release` it should have been dropped from. */ onDismiss: () => { this.holds.clear() this.hide() }, /* The side it landed on, said where a sheet outside can read it. The controller writes `data-placement` on the bubble and puts the arrow on the edge facing the trigger itself, so nothing here is needed to point it; this is for a sheet that draws by side, an entrance growing out of the trigger or an edge on the arrow's two outer sides, and that sheet is outside the shadow root, where no attribute can follow `::part()`. A part name because a tooltip is usually inside another component, where a document cannot reach its host's state. */ onPlace: (placement) => { const side = placement.split('-')[0] as Side this.toggleState('placed-bottom', side === 'bottom') this.toggleState('placed-left', side === 'left') this.toggleState('placed-right', side === 'right') this.toggleState('placed-top', side === 'top') this.side = side }, placement: () => this.placement, surface: () => this.partNamed('tooltip'), within: () => this.enclosure, })"
             },
             {
               "kind": "field",
@@ -99043,7 +99714,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -99397,6 +100068,22 @@
               "description": "While the tooltip shows.",
               "name": "open"
             },
+            {
+              "description": "While the bubble sits below its trigger.",
+              "name": "placed-bottom"
+            },
+            {
+              "description": "While the bubble sits to the left of its trigger.",
+              "name": "placed-left"
+            },
+            {
+              "description": "While the bubble sits to the right of its trigger.",
+              "name": "placed-right"
+            },
+            {
+              "description": "While the bubble sits above its trigger.",
+              "name": "placed-top"
+            },
             {
               "description": "Set a frame after the first render. Transitions are held off until it is, so an element never animates its own arrival.",
               "name": "settled",
@@ -99907,7 +100594,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -100326,6 +101013,22 @@
               "description": "The point, which the floating pass places against the target.",
               "name": "arrow"
             },
+            {
+              "description": "The point while the panel is below its target, so on the panel's top edge, beside `arrow`.",
+              "name": "arrow-bottom"
+            },
+            {
+              "description": "The point while the panel is left of its target, so on the panel's right edge, beside `arrow`.",
+              "name": "arrow-left"
+            },
+            {
+              "description": "The point while the panel is right of its target, so on the panel's left edge, beside `arrow`.",
+              "name": "arrow-right"
+            },
+            {
+              "description": "The point while the panel is above its target, so on the panel's bottom edge, beside `arrow`.",
+              "name": "arrow-top"
+            },
             {
               "description": "The control that goes to the previous stop.",
               "name": "back"
@@ -100463,6 +101166,15 @@
               "description": "Label for the control that leaves the tour.",
               "attribute": "skip-label"
             },
+            {
+              "kind": "field",
+              "name": "side",
+              "type": {
+                "text": "Side | undefined"
+              },
+              "privacy": "private",
+              "description": "The side of its target the panel last landed on, which names the arrow's\nsecond part. Undefined until the first placement, which the server never\nmakes."
+            },
             {
               "kind": "field",
               "name": "target",
@@ -100477,7 +101189,7 @@
               "name": "floating",
               "privacy": "private",
               "readonly": true,
-              "default": "new PopoverController(this, { anchor: () => this.target, arrow: () => this.partNamed('arrow'), /* Clear of the light, measured to the point of the arrow. The tour lights `spotlight-padding` around the target and the arrow reaches `n / sqrt(2)` past the panel's edge, and with no offset the panel sat on the target itself, over the lit margin, with its point 5.7px into what it pointed at. `offsetWidth` rather than a rect, which reports the rotated box. */ offset: () => { const tour = this.parentElement as (HTMLElement & { spotlightPadding?: number }) | null return ( (tour?.spotlightPadding ?? 0) + (this.partNamed('arrow')?.offsetWidth ?? 0) / Math.SQRT2 ) }, /* A press outside does not end a tour, and this is the decision rather than a stub. Everywhere else a surface closes on an outside press because pressing elsewhere means the reader is finished with it. Here elsewhere is the thing being pointed at: the scrim lets the pointer through on purpose, so the first press a reader makes on the button the tour is describing would be the press that ends the tour. Leaving is the skip control and `Escape`, both of which say so. */ onDismiss: () => {}, /* The side the panel landed on, which is not always `placement`: a stop with no room on the side it asked for is moved. The styled tier points its arrow from `data-placement` on the panel, inside this shadow root; a stylesheet outside cannot follow `::part(arrow)` with an attribute, and reading the stop's own `placement` instead pointed the arrow from the side that was asked for, or from below wherever React had set it as a property and there was no attribute to read. */ onPlace: (placement) => { const side = placement.split('-')[0] this.toggleState('placed-bottom', side === 'bottom') this.toggleState('placed-left', side === 'left') this.toggleState('placed-right', side === 'right') this.toggleState('placed-top', side === 'top') }, placement: () => this.placement, surface: () => this.partNamed('bubble'), })"
+              "default": "new PopoverController(this, { anchor: () => this.target, arrow: () => this.partNamed('arrow'), /* Clear of the light, measured to the point of the arrow. The tour lights `spotlight-padding` around the target and the arrow reaches `n / sqrt(2)` past the panel's edge, and with no offset the panel sat on the target itself, over the lit margin, with its point 5.7px into what it pointed at. `offsetWidth` rather than a rect, which reports the rotated box. */ offset: () => { const tour = this.parentElement as (HTMLElement & { spotlightPadding?: number }) | null return ( (tour?.spotlightPadding ?? 0) + (this.partNamed('arrow')?.offsetWidth ?? 0) / Math.SQRT2 ) }, /* A press outside does not end a tour, and this is the decision rather than a stub. Everywhere else a surface closes on an outside press because pressing elsewhere means the reader is finished with it. Here elsewhere is the thing being pointed at: the scrim lets the pointer through on purpose, so the first press a reader makes on the button the tour is describing would be the press that ends the tour. Leaving is the skip control and `Escape`, both of which say so. */ onDismiss: () => {}, /* The side the panel landed on, which is not always `placement`: a stop with no room on the side it asked for is moved. The floating pass puts the arrow on the edge facing the target itself; what a sheet draws by side is the edge on the arrow's two outer sides. The styled tier reads `data-placement` on the panel, inside this shadow root; a stylesheet outside cannot follow `::part(arrow)` with an attribute, and reading the stop's own `placement` instead drew from the side that was asked for, or from below wherever React had set it as a property and there was no attribute to read. */ onPlace: (placement) => { const side = placement.split('-')[0] as Side this.toggleState('placed-bottom', side === 'bottom') this.toggleState('placed-left', side === 'left') this.toggleState('placed-right', side === 'right') this.toggleState('placed-top', side === 'top') this.side = side }, placement: () => this.placement, surface: () => this.partNamed('bubble'), })"
             },
             {
               "kind": "method",
@@ -100619,7 +101331,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -101437,7 +102149,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -102338,7 +103050,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -102898,7 +103610,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -103458,7 +104170,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -104043,7 +104755,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -104676,7 +105388,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -106809,7 +107521,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",
@@ -107563,7 +108275,7 @@
                 "text": "Promise<void>"
               },
               "readonly": true,
-              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang race it\nagainst a timeout; `readyToMove` does exactly that.",
+              "description": "Resolves a frame after the first update, when transitions are switched on.\n\n**Named `whenSettled` because `settled` is taken four times over**, by\nImage, Splitter and both color controls, none of them about transitions. A\nbase class takes the name every subclass then cannot use.\n\n**The state it reports cannot be watched, which is why this exists.** A\ncustom state is deliberately not in the DOM, so there is no\n`MutationObserver` equivalent for `ElementInternals.states`. Await this\ninstead.\n\nSeparate from `updateComplete`, which resolves in the microtask checkpoint\nthe suppression is still active in. That is the distinction `readyToMove`\nin `@fusion.dev/helix-testing` is written around.\n\nIt never resolves where `requestAnimationFrame` never fires, which is a\nserver, a hidden tab, and happy-dom. Callers that must not hang there race\nit against a timeout, as the happy-dom harness's `readyToMove` does. The\nbrowser harness's awaits it alone: a browser always runs the frame, and a\ntimer beside it was a bet on how long one takes, which a busy run lost.",
               "default": "new Promise<void>((resolve) => { this.resolveSettled = resolve })",
               "inheritedFrom": {
                 "name": "HelixElement",