- nb.tsv covers every enforced key: menus, tl! literals, blend modes, brush sections, font weights, camera data and the generated preference labels (cargo xtask i18n-coverage: 1905/1905)
- Registered with the one|other plural rule; nb-NO, the macrolanguage no and Nynorsk nn-NO locales resolve to it
- License, attribution and localization docs updated
GrabCut's colour-model grids have few terminal links and short paths, where
BK beats the push/relabel solver the default dispatch picks for graphs of
4096+ nodes. Quick Selection keeps push/relabel.
24 MP synthetic, release, median of 3: Select Subject 480 -> 207 ms,
Object Selection 337 -> 140 ms, Quick Selection unchanged. IoU unchanged.
Reapply Linux Adwaita no-title winit feature and its lockfile,
keep integrated title bar as default, and port compatible tests/docs.
Preserve upstream's preferences UI refactor rather than restoring
its removed prefs_ui.rs file.
With Scroll All Windows on, or Shift held, a Hand, Space or middle-button
pan under an open dialog now moves the other shown documents to the same
relative location, as a pan on the free canvas does.
- The Stroke Color Picker on a shape whose stroke is off (none or 0 px)
previews the stroke its OK turns on, at the enabled width, even for the
colour it opened on (CodeRabbit).
- No Color's diagonal uses the theme's danger token, not a fixed red.
- docs/ui-parity.md: the history entry notes Recently Used Colors is a new
preference, so 12 of 182 do nothing (matching the row).
- The Fill/Stroke swatches clip their checker to the rounded corners, so
dark themes no longer show white specks there (#8741).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TYsCL5E2GM6hUabP9dmYAv
Upstream's strict-Clippy refactor removed glob imports and split the
Content-Aware Fill properties into caf_props_panel: the swatch title moved
there, and the frame loop calls crate::color_picker_ui::widget_tick.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TYsCL5E2GM6hUabP9dmYAv
step_damage narrows a layer whose only change is in its surfaces (pixels, mask, type,
smart-object and fill caches) to the bounding box of the tiles that differ: a merge walk of
the two tile maps, pointer compares only. A mask's tiles grow by its feather, all of them by
the layer's effects and the effects of the groups around it, never past the layer's whole
extents. A frame-spanning effect whose frame moved keeps the whole extents (#4053). Undo,
redo and History jumps get it for free: undoing a 60 px dab on a full-canvas layer redraws
its tiles, not the canvas, and so does a fill or filter on a small selection.
The painting commands' own damage rects go through one door, Session::hint_damage, which
intersects them with the edit's damage. Debug and test builds count a hint that misses a
changed tile in Session::damage_misses and log it; nothing panics. refresh_type_layers reports
step_damage instead of None.
Deleted: note_damage and layers_damage (identical to step_damage after layer.setProps and
layer.translate), note_inert_insert (step_damage already steps over inert layers, #1771),
and the engine's siblings_of they used.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Resolves the canvas.rs conflict by adopting the handle_primary_canvas_input
refactor and porting the menu-dismissal gates into it: the buttons starve,
the Crop press gate, the capture-press gate, the live-paint press gate, the
stroke started/moves/stopped/press-release gates, the zoom-hold on_canvas
gate, the release-click consume in the click handler, and the release
clear below the click handling. menu_dismiss_pass is computed once in
canvas_view and passed through CanvasToolInput.
Track-only rows that check they hit the whole-document-redraw path and
print the refreshed pixel count: undo of a brush dab on a 16-bit
15000x10000 full-canvas layer (reps capped at 5) and a fill of a 2 %
selection on a 24 MP full-canvas layer.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Only the outermost run_command writes the frame's command stage, so a nested
run (file.save -> layer.smartObjects.saveContents, actions.play -> its view
steps) isn't counted twice. ui_frame's UI moves into ui_body and finish_frame
runs after it on every path, including the screen picker's wait modal; the
pass before fonts are ready closes it too.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Track-only row, the slow-frame recorder's worst case. A 24 MP photo plus
three full-canvas blended copies (Multiply, Screen, Overlay); each sample
toggles the top copy and times one harness.step(), so every frame
recomposites all 24 MP on the CPU (the photo's tiles are never uniform,
which keeps the reduced render's one-colour shortcut out of the way).
The row errors unless each frame recorded a CPU refresh covering at least
95 % of the canvas, so it can't silently take a fast path. 10 samples.
- The PNG encoder gets a fallibly reserved copy of the composite's pixels instead of a
clone, and a full-size composite that cannot be copied or encoded is left out of the
ZIP or directory bundle and out of its manifest; the save goes on without it.
- The ZIP size limit for keeping the full-size composite counts the central directory
and end record as well as the entries.
- `pcraftFullComposite` that is not true, false or null is a bad parameter of
`file.saveACopy`, the other engine saves and the automation `doc.save`.
- MCP `doc_save` in bridge mode refuses `pcraftFullComposite: false` as well as `true`.
The app's `app.save` does not take the option.
Add gradient_midpoints() which reads Mdpn (0–100) from each color stop
descriptor, converts to the engine's 0..1 range, and returns one value
per segment (n-1 for n stops; the last stop's Mdpn is irrelevant). Returns
None when all segments are at the default (0.5) so the key is omitted.
gradient_layer_set_params now inserts "midpoints" when any segment differs
from the default, passing it to gradient.fill.set which accepts a flat
float array at that key.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Fix 2: gradient_color_stops now returns None when fewer than 2 color stops
are present, preventing a partial broken gradient layer from being created
from a malformed descriptor.
Fix 3: gradient_layer_set_params now always writes explicit values for
reverse (true/false), dither (true/false), scale (value or 100.0), offset
([h, v] always, defaulting to [0, 0]), and transparency stops (always
included, even if all opacities are 100).
Fix 1 (midpoints) not applied: gradient.fill.set does not accept midpoints
as a third element on individual stops. It accepts them via a separate
"midpoints" array parameter — that conversion is deferred.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The flick-pan frame and the Hand pan read response.dragged() directly,
before menu_dismiss_press is computed, and live in the hand_pan branch
where the zeroed buttons never arrive: with the menu dismissed and the
dismissing button still down, holding Space (or the selected Hand tool)
and dragging panned the view. The gate is now computed up front and
guards both pan sites (CodeRabbit, #9841).
Regression test a_held_menu_dismissal_press_cannot_pan_the_view fails
on the previous commit and passes now.
crop_ui.rs moved to tools/; the PR's builders replace main's Clippy extractions of the same code; #6723's Color Lookup save carried into the moved encoder.
Addresses the two remaining CodeRabbit findings on PR #9841:
- The dismissal flag is cleared below the click handling, so a later-frame
release over the canvas is consumed by the flag check first (clearing it
before let that release zoom). canvas_tool_menu::show does not arm the flag
when the frame already contains the release, so a fast same-frame click
cannot eat the next ordinary click.
- The blocked gesture now persists in canvas_tool_menu_dismissed until its
release instead of being re-derived from the open menu each frame, so a held
dismissal press stays blocked after the menu closes. Navigation is blocked
too: the animated zoom hold cannot arm on it, and zoom drags, Rotate View
and other button-driven gestures are starved.
Regression tests (fail on the previous commit, pass now):
- dismissing_the_view_menu_consumes_the_whole_gesture
- a_held_menu_dismissal_press_never_arms_the_zoom_hold