Give undo and the card picker's buttons a fingertip to land on

The first thing the repaired audit found, run on a real touch context at 44px: three
controls a dealer taps on a phone that are too small to tap.

    .actionUndo             145x30   the "I recorded that wrong" pill
    #cardPickerClearBtn      50x40   wipes a player's cards
    #cardPickerCloseBtn      50x40   closes the picker

Undo is the worst of them. It is the control a dealer reaches for having just entered
the wrong action with five people waiting, and it was fourteen pixels short of a
fingertip. The picker's two are short because the touch rule that fixes every other
small button is scoped to .panelBody, and the picker is a sheet — it never reached
them. One of them is destructive.

All three now sit at 44px on touch. Nothing changes on a pointer device.

I tried an invisible hit-area pad on undo first, to avoid spending bar height, and the
measurement said it did nothing: the pill clips its overflow to keep a long label on
one line, and giving it a positioning context made it clip the pad too. Growing the
box is the honest fix, and it turned out to cost nothing anyway — the bar goes from
138px to 152px and the felt stays at 366px, because the felt had slack against its
220px floor rather than sitting on it as I had assumed.

Co-Authored-By: Claude Opus 5 <[email protected]>
Claude-Session: https://claude.ai/code/session_015bnp7xKUPFFSpwkYNFaPgb

JoshHeaps.Net/wwwroot/css/casino/holdem.css

@@ -3138,7 +3138,18 @@ body.hasActionBar #toastHost {
     }
 
     .panelBody .holdemBtnSmall,
-    .handHistoryActions .holdemBtn {
+    .handHistoryActions .holdemBtn,
+    /* The card picker's Clear and Done sat at 40px: the touch rule was scoped to
+       .panelBody and the picker is a sheet, so it never reached them. Clear wipes a
+       player's cards, which is not a thing to make a dealer aim at. */
+    .cardPickerHeader .holdemBtn,
+    /* Undo was 30px, and it is the control a dealer reaches for having just recorded
+       the wrong thing with five people waiting — the worst place to be fourteen pixels
+       short of a fingertip. Growing the box costs bar height, which comes out of the
+       felt, but measured at 390x844 the felt is 366px against a 220px floor, so there
+       is room for it. A hit-area pad was tried first and does not work here: the pill
+       clips its overflow to keep a long label on one line, which clips the pad too. */
+    .actionUndo {
         min-height: 44px;
     }
 
@@ -3525,7 +3536,9 @@ textarea:focus-visible {
 
     .panelBody .holdemBtnSmall,
     .handHistoryActions .holdemBtn,
-    .toastAction {
+    .toastAction,
+    .cardPickerHeader .holdemBtn,
+    .actionUndo {
         min-height: 44px;
     }
 

tests/holdem/uiAudit.js

@@ -93,10 +93,44 @@ const LayoutAudit = {
         return false;
     },
 
+    // The modal that is currently on top, if any.
+    //
+    // Every check below asks "is this thing reachable", and while a sheet or dialog is
+    // open the answer for everything behind it is legitimately no — that is what a
+    // modal is for. Without knowing that, the audit reported the pot wizard covering
+    // the seats as ten faults and the rulings sheet as seven, which is the noise that
+    // made it useless in exactly the states nobody had audited. Now it scopes to what
+    // is actually on top and says so.
+    MODAL_SELECTOR: "#potWizardOverlay, #rulingsOverlay, #seatMenuOverlay, #cardPickerOverlay, " +
+        "#confirmOverlay, #addChipsOverlay, #seatEditorOverlay",
+
+    openModal() {
+        const open = [...document.querySelectorAll(this.MODAL_SELECTOR)]
+            .filter(el => !el.hidden && getComputedStyle(el).display !== "none");
+        // Last in the DOM wins, which is how they stack.
+        return open.length > 0 ? open[open.length - 1] : null;
+    },
+
+    // A seat sheet on a phone isn't an overlay element, it's a body class — but it
+    // covers the felt just the same.
+    modalScope() {
+        const overlay = this.openModal();
+        if (overlay) return overlay;
+        if (document.body.classList.contains("seatSheetOpen")) {
+            return document.getElementById("seatEditorPanel") || null;
+        }
+        return null;
+    },
+
     occlusion(selector = "#actionBar button, #holdemHeader button, .seat") {
         const offenders = [];
         const warnings = [];
-        [...document.querySelectorAll(selector)].forEach(el => {
+        // While something is open on top, only what is inside it can be judged. The
+        // rest of the page is meant to be unreachable.
+        const scope = this.modalScope();
+        const candidates = [...document.querySelectorAll(selector)]
+            .filter(el => !scope || scope.contains(el));
+        candidates.forEach(el => {
             const rect = el.getBoundingClientRect();
             if (rect.width === 0 || rect.height === 0) return;
 
@@ -504,10 +538,25 @@ const LayoutAudit = {
         if (!enforceOnPointerDevices && matchMedia("(hover: hover)").matches) {
             return { name: `touch targets at least ${minimum}px`, offenders, skipped: "pointer device" };
         }
+        const unmeasured = [];
         this.visibleControls()
             .filter(el => el.type !== "color")
             .filter(el => !(exempt && el.matches(exempt)))
             .forEach(el => {
+                // A control inside a scrolling sheet can sit below its container's
+                // visible area while still having a viewport rect. elementFromPoint at
+                // its centre then returns the sheet, effectiveTarget gives up, and the
+                // check used to pass it in silence — which is how "Confirm Winner(s)",
+                // the button that pays the pot out, went unmeasured for 36 rounds.
+                // Bring it into view before probing.
+                const scroller = el.closest("[style*='overflow'], .modalBody, #potWizardBody, #rulingsSheet, #seatEditorPanel");
+                if (scroller && scroller.scrollHeight > scroller.clientHeight) {
+                    const box = scroller.getBoundingClientRect();
+                    const own = el.getBoundingClientRect();
+                    if (own.bottom > box.bottom || own.top < box.top) {
+                        el.scrollIntoView({ block: "center" });
+                    }
+                }
                 const rect = el.getBoundingClientRect();
                 // Drawn big enough is big enough — no need to probe, and probing wide
                 // elements is where the cap above used to produce nonsense. Seats sit
@@ -515,15 +564,35 @@ const LayoutAudit = {
                 // measuring their reach would report the crowding as a fault.
                 if (rect.width >= minimum - 1 && rect.height >= minimum - 1) return;
 
+                // A checkbox inside a label is tapped by tapping the label, so the
+                // label is the target. Measured rather than probed, because probing
+                // fails on anything scrolled off the viewport — which is how a 24px
+                // checkbox with a 335x44 label around it came up as unprobeable
+                // instead of as the perfectly good target it is.
+                const label = el.closest("label");
+                if (label) {
+                    const lr = label.getBoundingClientRect();
+                    if (lr.width >= minimum - 1 && lr.height >= minimum - 1) return;
+                }
+
                 const size = this.effectiveTarget(el);
-                if (!size) return;
+                // Still can't measure it. That is not a pass — it is the check failing
+                // to run, and quietly returning here is exactly the kind of exemption
+                // that hides real faults. Say so instead.
+                if (!size) {
+                    unmeasured.push(
+                        `${this.describe(el)} :: could not be probed` +
+                        ` (drawn ${Math.round(rect.width)}x${Math.round(rect.height)})`
+                    );
+                    return;
+                }
                 if (size.width >= minimum - 1 && size.height >= minimum - 1) return;
                 offenders.push(
                     `${this.describe(el)} :: hittable ${size.width}x${size.height}` +
                     ` (drawn ${Math.round(rect.width)}x${Math.round(rect.height)})`
                 );
             });
-        return { name: `touch targets at least ${minimum}px`, offenders };
+        return { name: `touch targets at least ${minimum}px`, offenders, unmeasured };
     },
 
     // Tapping the middle of a seat must give you that seat.
@@ -537,6 +606,13 @@ const LayoutAudit = {
     // Reported separately from occlusion because the cause is different: not an
     // overlay on top of the table, but two pieces of the table fighting over a tap.
     seatOwnership() {
+        // Seats behind an open sheet aren't fighting each other for taps — nothing is
+        // meant to reach them at all. Asking the question there produced six offenders
+        // per modal, every one of them the modal doing its job.
+        if (this.modalScope()) {
+            return { name: "a seat's centre belongs to that seat", offenders: [],
+                skipped: "a sheet is open over the table" };
+        }
         const seats = [...document.querySelectorAll(".seat")];
         const offenders = [];
 
@@ -617,7 +693,9 @@ const LayoutAudit = {
             viewport: `${innerWidth}x${innerHeight}`,
             failed: checks.filter(c => c.offenders.length > 0),
             // Things worth knowing that aren't faults — see occlusion() on toasts.
-            warnings: checks.flatMap(c => c.warnings || []),
+            // Controls a check couldn't probe go here too: not a fault, but not a pass
+            // either, and a silent skip is how faults hide.
+            warnings: checks.flatMap(c => [...(c.warnings || []), ...(c.unmeasured || [])]),
             passed: checks.filter(c => c.offenders.length === 0).map(c => c.name),
             clean: checks.every(c => c.offenders.length === 0),
         };