fitCoverRect method

Rect fitCoverRect(Rect baseRect, Offset touchGlobalPosition, { double pullFactor = 0.4, })

Computes the fit-cover end rect for a double-tap on a displayed (at-base) view. Scales the rect from BoxFit.contain to BoxFit.cover (so it fills the display), then pans toward touchGlobalPosition on the overflowing axis (clamped to cover bounds).

this is the displayed container, baseRect is the at-base rect (BoxFit.contain inside this).

pullFactor ∈ 0, 1 selects the panning behavior:

• 0 → Apple-style: touched point stays under finger (no edge attraction).
• 1 → ref-style: max edge attraction (taps near the edge snap the rect's matching edge against the container edge).
• In between: blended. Clamp is always applied so the rect can't slide past cover bounds (would expose container background).

Implementation

Rect fitCoverRect(
  Rect baseRect,
  Offset touchGlobalPosition, {
  double pullFactor = 0.4,
}) {
  final coverScale = area / baseRect.area;
  final coverWidth = baseRect.width * coverScale;
  final coverHeight = baseRect.height * coverScale;
  // "fit by width" = base spans the full container width (height has slack);
  // in cover state, width overflows, so we pan along x. Height-fit is the
  // mirror case.
  final fitByW = (baseRect.width - width).abs() < 0.5;
  final touchPoint = center - touchGlobalPosition;
  // k bridges Apple (coverScale - 1) and ref (coverScale).
  final k = (coverScale - 1) + pullFactor;
  final dxLimit = center.dx * (coverScale - 1);
  final dyLimit = center.dy * (coverScale - 1);
  return Rect.fromCenter(
    center: center.translate(
      fitByW ? (touchPoint.dx * k).clamp(-dxLimit, dxLimit) : 0,
      fitByW ? 0 : (touchPoint.dy * k).clamp(-dyLimit, dyLimit),
    ),
    width: coverWidth,
    height: coverHeight,
  );
}