feat(pen): extensible brush model (4 brushes)
All checks were successful
CI / Windows build (push) Successful in 14m54s

Replace the 2-tool ink system with a data-driven, Krita-style
BrushProfile (lib/editor/engine/brush.dart). Adding a brush is a
const map entry, not render-path branching.

Four presets from the rnote/krita spec:
- fountain pen: quadratic (p^2) pressure, wide dynamic width
- ballpoint:    near-constant width (thinning 0.15)
- highlighter:  flat width, square caps
- pencil:       sqrt(p) pressure, moderate width

Pressure is pre-warped per brush via PressureCurve(gamma) before
perfect_freehand; geometry fields (thinning/streamline/smoothing/
caps) flow through the shared stroke recipe so the PDF overlay and
the note/slide PenCanvas both honor the brush. Brush kind is now
persisted on the stroke model. Picker added to all three toolbars.

Opacity/multiply and pencil grain are carried as data but not yet
composited (TODO brush-opacity / brush-texture); this increment is
width + pressure-curve differentiation. analyze clean, 283 tests.
This commit is contained in:
2026-06-24 11:13:43 +08:00
parent 45a8931b64
commit 0feca74278
21 changed files with 805 additions and 66 deletions

View File

@@ -14,6 +14,7 @@ import 'package:flutter_test/flutter_test.dart';
import 'package:badnote/editor/canvas/ink_painters.dart';
import 'package:badnote/editor/canvas/pen_stroke.dart';
import 'package:badnote/editor/engine/brush.dart';
import 'package:badnote/editor/engine/stroke_geometry.dart';
import 'package:badnote/editor/engine/stroke_model.dart';
@@ -44,22 +45,27 @@ void main() {
expect(byDefault.getBounds(), explicit.getBounds());
});
test('thinning actually affects the outline (not hardcoded/ignored)', () {
// NB: the bounding box is thinning-INVARIANT here because
// perfect_freehand's round end-caps are drawn at the full `size`; only
// the mid-section width tracks pressure×thinning. So we compare the
// outline PERIMETER (sum of contour lengths), which does reflect the
// pinched middle.
test('brush thinning actually affects the outline (param is wired)', () {
// After the brush-engine rebuild each brush owns its perfect_freehand
// thinning (spec §4): fountainPen = 0.9 (pressure-modulated, pinched
// middle) vs highlighter = 0.0 (constant full width). The bounding box is
// thinning-INVARIANT (round caps at full size), so compare the outline
// PERIMETER, which reflects the pinched middle.
double perimeter(Path p) =>
p.computeMetrics().fold(0.0, (sum, m) => sum + m.length);
final pen = _pressuredPen();
final strong = perimeter(
buildStrokePath(pen, size, isComplete: true, thinning: 0.85));
final none = perimeter(
buildStrokePath(pen, size, isComplete: true, thinning: 0.0));
// Constant width (0.0) vs pressure-thinning (0.85) must differ measurably.
final pressured = _pressuredPen(); // fountainPen brush, thinning 0.9
final flat = PenStroke(
points: _pressuredPen().points,
color: 0xFF000000,
width: 0.01,
kind: PenStrokeKind.highlighter,
brush: BrushKind.highlighter, // highlighter brush, thinning 0.0
);
final strong =
perimeter(buildStrokePath(pressured, size, isComplete: true));
final none = perimeter(buildStrokePath(flat, size, isComplete: true));
expect((strong - none).abs(), greaterThan(1.0),
reason: 'thinning had no effect on the outline — it is not wired');
reason: 'brush thinning had no effect on the outline — not wired');
});
test('screen and export builders agree for the same stroke + thinning', () {