Description: Remove unused shader helpers left after refactor and restore probe visibility for grid vs watermark cost (without changing current visual behavior).
Acceptance: No unused-element warnings in touched files; probe still distinguishes relevant background costs.
Description: Remove unused shader helpers left after refactor and restore probe visibility for grid vs watermark cost (without changing current visual behavior).
Acceptance: No unused-element warnings in touched files; probe still distinguishes relevant background costs.
First half done on development (2026-09-05, uncommitted at time of writing). Second half still open, so leaving this issue open on the reduced scope.
Dead code removed._screenRepeatShaderMatrix is gone from background.dart - it had zero callers and was the last survivor of the rasterised-tile grid approach the fragment shader replaced.
Went further than the title while in there, since it was the same class of thing in the same paint path. Removed from canvas_painter.dart:
_v2LineTileImages / _v2UnderlayTileImages / _v2StationTopTileImages and their three *InFlight sets - all declared, cleared in the reset path, never populated or read
_v2StaticCompositeKey / _v2StaticCompositePicture / _v2StaticCompositeGuardRect - same shape _qtTileImages is NOT dead and was left alone; that is the live qt pyramid cache. flutter analyze clean, full suite green.
Still open: restore the probe split metrics. Grid vs watermark cost is still not separable in the probe output, which is the more useful half of this issue and the reason it stays open. Related to the perf HUD trustworthiness question filed today.
First half done on development (2026-09-05, uncommitted at time of writing). Second half still open, so leaving this issue open on the reduced scope.
**Dead code removed.** `_screenRepeatShaderMatrix` is gone from `background.dart` - it had zero callers and was the last survivor of the rasterised-tile grid approach the fragment shader replaced.
Went further than the title while in there, since it was the same class of thing in the same paint path. Removed from `canvas_painter.dart`:
- `_v2LineTileImages` / `_v2UnderlayTileImages` / `_v2StationTopTileImages` and their three `*InFlight` sets - all declared, cleared in the reset path, never populated or read
- - `_v2StaticCompositeKey` / `_v2StaticCompositePicture` / `_v2StaticCompositeGuardRect` - same shape
`_qtTileImages` is NOT dead and was left alone; that is the live qt pyramid cache. `flutter analyze` clean, full suite green.
**Still open: restore the probe split metrics.** Grid vs watermark cost is still not separable in the probe output, which is the more useful half of this issue and the reason it stays open. Related to the perf HUD trustworthiness question filed today.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Description: Remove unused shader helpers left after refactor and restore probe visibility for grid vs watermark cost (without changing current visual behavior).
Acceptance: No unused-element warnings in touched files; probe still distinguishes relevant background costs.
First half done on development (2026-09-05, uncommitted at time of writing). Second half still open, so leaving this issue open on the reduced scope.
Dead code removed.
_screenRepeatShaderMatrixis gone frombackground.dart- it had zero callers and was the last survivor of the rasterised-tile grid approach the fragment shader replaced.Went further than the title while in there, since it was the same class of thing in the same paint path. Removed from
canvas_painter.dart:_v2LineTileImages/_v2UnderlayTileImages/_v2StationTopTileImagesand their three*InFlightsets - all declared, cleared in the reset path, never populated or read_v2StaticCompositeKey/_v2StaticCompositePicture/_v2StaticCompositeGuardRect- same shape_qtTileImagesis NOT dead and was left alone; that is the live qt pyramid cache.flutter analyzeclean, full suite green.Still open: restore the probe split metrics. Grid vs watermark cost is still not separable in the probe output, which is the more useful half of this issue and the reason it stays open. Related to the perf HUD trustworthiness question filed today.