Hypothesis, from WebKit's fixedContainerEdges: Safari probes
one point per edge — 4px in from the edge, horizontally
centered — walks up to the nearest fixed/sticky
ancestor, and paints its chrome opaque only when that element is roughly
full viewport width (≥ 90%). DOM depth should be irrelevant.
Open on iPhone Safari, scroll until the bars pin, then work the toggles.
The panel shows the hypothesis's prediction for the current combo — if
Safari ever disagrees, the hypothesis is wrong.
The current config is mirrored into the URL, so a reload starts in that
state — handy once the opaque chrome gets stuck and won't reset in place.
Gap 5px → 4px: flips translucent → opaque. The probe sits exactly 4px in.
Width 100% → 85%: flips opaque → translucent. The probe still hits the bar, but it fails the 0.9 ratio.
48% left-aligned: translucent even at 0px gap. The centered probe never touches the bar.
Depth 1 → 4 wrappers: changes nothing, in any combo.
Decoy on: back to translucent even with a fully qualifying bar behind it.