The rounding overshoot fixed in 6d4df2b9, through the presenter's own frames rather than a direct call: a tall frame presented fit-to-width at eleven fractional widths, and a still view keeps its detail texture for four more frames. Before 6d4df2b9 it failed at width 1555.5 (a new texture on every frame).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove and wrote
The page rect of the view, converted back from the screen, can stand a hair outside the frame. The cached detail is always inside the frame, so it never covered that view, and a still view rendered a new detail every frame. detail_texture now clips the view to the frame first. Test: present::tests::a_view_a_rounding_error_past_the_frame_keeps_its_detail (failed before: a new texture on the second call).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove and wrote
Both textures start from a transparent pixmap, so a translucent layer was counted twice where the detail covered it. The whole-frame texture is now drawn only around the detail rect (up to four clipped strips), then the detail.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
Rust 1.100 marks `ExitCode` as must-use and extends `needless_bool`,
so `cargo clippy --workspace --all-targets -- -D warnings` fails on main:
- `crates/ui-egui/src/shortcuts.rs`: the end of `key_event` returns the
condition directly instead of `if ... { return true; } false`.
- `apps/figcraft-cli/src/files.rs`: three script tests now assert the exit
code that `run_script` returns (`SUCCESS` for the two scripts that run,
1 for the script whose step fails) instead of dropping it.