Review follow-up (CodeRabbit): a linear gradient expands to a mesh with a
column at each stop, and the mesh mixes linearly between columns, so a
Perceptual blue -> yellow came out as the grey RGB mix in the middle.
`breaks` now also takes a column at each stop the Perceptual expansion adds
(Gradient::expanded), so the mesh follows the OKLab curve; Classic is
unchanged. the_mesh_option_makes_a_mesh_that_paints_the_gradient gains a
Perceptual case; without the change 1812 pixels differ from the gradient.
Review follow-up (CodeRabbit): the test's doc comment quoted a phrase from
the public help page. AGENTS.md allows feature names and a cited URL, not
copied wording; the comment now describes the behaviour itself and keeps
the URL.
Illustrator 30.0's Method choice in the Gradient panel
(https://helpx.adobe.com/illustrator/desktop/paint-and-fill/create-and-edit-gradients/control-perceptual-interpolation-in-gradients.html).
- paint.editGradient takes `method: classic|perceptual` (one undo step);
paint.setFill/setStroke take it in `gradient`, and gradient params (swatches,
copied paints) carry it when Perceptual. Anything else is refused without
editing. One edit can set `method` and `dither` together.
- The Gradient panel shows a Method menu (Classic, Perceptual) for linear and
radial gradients, dimmed without a gradient; the Dither box stays under the
stops.
- i18n: "Perceptual" where a complete catalog lacked it, and the `method`
error in all 24 catalogs.
- Docs: G13 (estimate 14–24 h → 13–23 h, after Dither's 15–25 h → 14–24 h),
target-app-parity (row and evidence), ui-parity, mcp.md, plugins.md.
Illustrator 30.0's gradient Method, from the public documentation
(https://helpx.adobe.com/illustrator/desktop/paint-and-fill/create-and-edit-gradients/control-perceptual-interpolation-in-gradients.html):
Classic or Perceptual; perceptual interpolation matches colour transitions to
how the eye perceives differences between colours, and perceptual gradients
are rasterised when saved as PDF or SVG.
- Gradient gains `method` (Classic | Perceptual; serde default Classic,
written only when Perceptual, so older files open as Classic and Classic
files are written as before). The colour crate's OKLab helpers open up for it.
- Perceptual mixes neighbouring stop colours in OKLab (sample, and
sample_keep_model for the stroke meshes, Expand and blends); opacity stays
linear. Renderers that mix linearly get stops on the perceptual curve,
placed by halving each half segment where a straight line strays more than
0.4 of a colour level (a nearly straight segment needs none). Classic
gradients expand exactly as before.
- draws_as_pixels includes Perceptual gradients, so SVG and PDF write a
Perceptual fill or stroke as an image at the document's raster effects
resolution, clipped to what it paints. A gradient along or across a stroke
stays slices (PDF no longer draws it as a within-stroke image).
- With Dither: a dithered Perceptual fill is the perceptual mix with dither's
noise (the canvas's dithered grids are keyed by method too); SVG and PDF
write it as an image with the noise, and a dithered Perceptual stroke as an
image without it, as the canvas draws it (GradientPaint::pixels, SVG's
raster_paint and PDF's area are told whether they paint a fill).
Review follow-up (CodeRabbit): the refusal skipped any request carrying a
`swatch` key, so `"swatch": null` with a dithered gradient got through.
Only a named swatch (a gradient swatch keeps its own dither) is exempt now.
The dither test sends `"swatch": null`; it fails without the change.
Review follow-up (CodeRabbit): a gradient drawn as pixels was sized from its
area in the space it is drawn in, so type scaled 4 times got a quarter of the
raster effects resolution; the image is now sized from the area on the page,
within the same pixel caps. And type whose fill is drawn as art (a dithered
or freeform gradient's image, a pattern) went out as outlines only; its
characters are now also written as invisible text (rendering mode 3), as
unpainted type already is, so they can be selected and searched.
Review follow-up (CodeRabbit): paint.setStroke kept `dither: true` from its
gradient, and a stroke can hold dither anyway (swapped from a fill). Outline
Stroke and appearance expansion copied the stroke's paint into the outline's
fill, which then drew dithered although the stroke never did. The outline's
fill now goes without dither (Paint::stroke_as_fill), and paint.setStroke
refuses a gradient with `dither: true`, as paint.editGradient does.
Review follow-up (CodeRabbit): a dithered fill skipped set_paint's guard for
a gradient without stops and sampled a transparent grid to draw nothing. It
now takes the same no-fill path.
Review follow-up (CodeRabbit): a large dithered gradient mostly off-screen
sampled its whole grid (up to 4096 x 4096 cells, 64 MiB) to draw a small
visible part, and did it again on each zoom step. The renderer now samples
only the cells the render target can show, a cell more each way for the
bilinear mix, snapped out to 256-cell tiles so a pan within them reuses the
grid. Cells keep their place in the whole grid, so the gradient and its
noise are the same pixels as before (GradientPaint::pixels_within).
Illustrator 30.0's gradient Dither, from the public documentation
(https://helpx.adobe.com/illustrator/desktop/paint-and-fill/create-and-edit-gradients/control-dither-in-gradients.html):
a Dither checkbox in the Gradient panel; dither converts the gradient into
pixels with subtle noise, works with gradient fills only, and follows the
Document Raster Effects resolution.
- Gradient gains `dither` (serde default false, written only when on, so
older files open unchanged and new files keep it).
- The renderer draws a dithered linear or radial fill as an image paint: the
gradient sampled a cell per device pixel or finer (quarter-octave steps,
cached per thread), each channel given seeded triangular noise of up to a
level before rounding. Same pixels every run; strokes and undithered fills
draw exactly as before. PNG and other raster exports go through it.
- SVG and PDF write a dithered fill as an image at the document's raster
effects resolution, clipped to the shape (the freeform-gradient path);
characters' fills stay vector gradients in SVG, with a warning.
- paint.editGradient and paint.setFill take `dither`; strokes refuse it.
The Gradient panel's Dither box is dimmed behind the Stroke proxy.
- i18n rows in all 24 catalogs; docs: G13, target-app-parity, ui-parity,
mcp.md, plugins.md.
Shared groundwork for gradient options SVG and PDF can't express (Illustrator
30's dither and perceptual gradients): GradientPaint::pixels samples a placed
linear or radial gradient on a grid (Ramp: its stops, midpoints as stops, mixed
as the renderer mixes them), and draws_as_pixels says which gradients are drawn
that way (freeform ones, so far). The SVG and PDF writers' freeform-gradient
image path takes any such gradient, a linear or radial one at the document's
raster effects resolution. No change in output yet.
Review follow-up (CodeRabbit): a line that wasn't valid JSON was reported
by its line only, while every other --commands error names the run's step.
It now says `step N (line L)`, N counted with the --cmd steps before it.
run_refuses_malformed_commands_naming_the_step expects `step 2 (line 3)`;
it fails without the change.
Review follow-up (CodeRabbit): the two older G2 bullets counting 46 and 49
filters were superseded by Conté Crayon (57/57) and are folded into one
current bullet; the G2 row in the ranked table and the entry's estimate now
both say 4–6 h, as ROADMAP.md and roadmap.md do. ROADMAP.md's revision
history gains its 2026-10-22 row.
Review follow-up (CodeRabbit): the blank-name check ran on the raw name, so
a name made only of characters XML can't hold (control characters) wrote an
empty <title>. The name is sanitised first, then the fallback applies. The
title test covers such a name; it fails without the change.
The SVG writer always wrote the document's title in <title>
(crates/svg/src/export.rs), so every file of Export As with Use
Artboards, Export for Screens or an `artboard`/`range` export was titled
after the document ("Icons"), whatever its artboard was called. The
writer's `artboard` option can't tell those apart from Save: Save writes
the first artboard too, standing for the whole document.
ExportOptions has a new `artboard_title` (set by the caller, as
`artboard` is, not deserialized): the file is that one artboard, exported
on its own, and its <title> is the artboard's name (XML-escaped as
before), or the document's title when the artboard has no name. The
engine's SVG encoder (fileio/svg.rs) sets it exactly when the export
names its artboards, which is also when a file holds only that
artboard's art (#550): Export As per artboard, document.export /
serialize with `artboard` or `range`, MCP export, and Export for Screens'
artboards, assets, layers and selections (each written as an artboard
named after it). Save and Save As keep the document's title. Export for
Screens' Full Document draws the art bounds as one artboard, which is now
named after the document, so its SVG keeps the document's title too.
docs/file-format-parity.md's SVG row says so.
Tests: title_control_chars::an_artboard_exported_alone_is_titled_after_it
(crates/svg: two named artboards, the second named with markup;
exporting either alone gives its escaped name, the default and an
artboard without artboard_title give the document's title, an unnamed
artboard falls back to it) and
tests_svgenc::an_svg_of_one_artboard_is_titled_after_it (the engine:
`artboard` 0 and 1, useArtboards range all, Save's SVG, Export for
Screens per artboard, an asset and Full Document). Both fail without the
change (the title was "Icons") and pass with it.
Review follow-up (CodeRabbit): step_value clamped once after all of a
frame's steps, so at the lower bound a step down then a step up in the
same frame left the field at the bound. Each step is now clamped as it is
applied. The Stroke panel test steps down and up in one frame at each
bound and expects one step above it; it fails without the change.
Wheel notches and the arrow keys over a focused number field step the
value its text shows (widgets::step_value), and nothing there knew the
field's bounds: the Stroke panel's Weight went on to -3 pt and its Limit
to 0 x in the field, while the panel only clamped what it applied (the
spinner's min, Limit's clamp to 1..500), so the next step went on from
the out-of-range text. A label scrub (scrub::field) had the same gap.
step_value and scrub::field now take the field's bounds and stop at
them, so the field shows where they stopped and the value applied is
the same. The fields that have bounds pass them:
- spin_field / spin_field_auto / spin_plain: their `min` (the spinner
and a typed value already stopped there);
- range_field: its range; it now also takes a value that may be blank
(None), as mixed_field does;
- opt_field (dash and gap lengths): 0;
- the Revolve dialog's values: their range.
Fields that clamped only what they applied now say their range with
range_field instead: the Stroke panel's Limit (1..500), the Transparency
and Properties opacity (0..100), the Transform panel's polygon sides,
and the dialogs' bounded counts, percentages and lengths (Blend Options,
Envelope, Flattener, Liquify, Freehand, Live Paint, Path ops, Perspective
Grid, PNG ppi, Print, Save for Web, Shape Builder, Slices, Swatch
Options, Symbol options, Justification, Dimension, PDF page, the effect
dialogs' number rows). docs/ui-parity.md's number-field row says so.
Test: panels::stroke::tests::the_wheel_stops_weight_at_zero_and_limit_at_one
(the Stroke panel in a headless frame; focus Weight at 1 pt and turn the
wheel down four notches: the field shows 0 pt and the stroke is 0 pt;
the same with Limit at 2 x: 1 x and a limit of 1). It fails without the
change (the field showed -3 pt) and passes with it; the other numeric
field, scrub and dialog tests still pass.
The MCP drawing tools ran the creating command (path.create, shape.*)
and only then paint.setFill, paint.setStroke, stroke.set,
object.setProps and effect.apply, each as a command of its own.
check_draw_options checked dash, cap, join, opacity, name and effects up
front, but not the paints: a fill or stroke the engine can't read
(stroke: "currentColor"), or a swatch the document lacks, failed after
the object was made, so the call returned an error and left the path,
its undo step and its SVG node behind. Each styling command was also an
undo step of its own.
The tools now send the creating command and the styling commands as one
command.batch (crates/mcp/src/tools.rs, draw): the engine rolls the
whole batch back when a step fails, so a refused paint leaves no object,
no undo step and nothing in the SVG, and a good call is one undo step
named as the creating command names its own (Pen, Rectangle, …). The
styling steps act on the new object, which the creating command selects,
with item: null (its whole appearance), as the ids did before. The
reply gives the failed step's own error, without the batch's prefix.
docs/mcp.md says so.
Test: tests_drawopts::a_bad_paint_draws_nothing (invalid fill, stroke,
a missing swatch and a non-numeric strokeWidth on draw_path and
draw_shape: error, node count, undo history and SVG export unchanged;
a good call adds one undo step named Pen / Rectangle). It fails without
the change (draw_path with stroke "currentColor" drew a path) and passes
with it; the other draw option and command_batch tests still pass.
The Layers list scrolls in its own scroll area, which kept egui's floating
scroll bar when #3093 gave the panel bodies a bar in a column of their own.
Hovering the list showed the bar over the right edge of the rows, over the
target circles and the selection squares the user clicks to select a layer.
The list now uses the menus' slim solid bar (list_scroll_style) in its own
column, and the rows end beside it. The list knows its total height before
it lays the rows out, so the bar is shown exactly when the rows overflow
(decided up front): the rows never shift sideways a frame after the bar
appears.
Test: dock::tests::the_layers_scroll_bar_covers_no_row (40 layers in a
1440x600 window: no row's accessibility rect intersects the bar). It fails
without the change and passes with it. The Layers tests that pin row
positions across frames still pass. Checked live: the bar sits beside the
rows.
With the pointer resting over art while the view moves (trackpad or wheel
scrolling, pinch zoom, a Hand/space pan), the document point under the
pointer changes every frame, and the canvas hit-tested the art there on
each one, up to three times: Object Highlighting (canvas.rs
hover_highlight -> selection_hit_at) and the Selection tool's cursor
(select.rs cursor: hit_test, and with Snap to Point the nearest-anchor
search). Over a detailed engraving in a clip group (the reporter's
PhotoCraft icon: one path of tens of thousands of segments) each of those
costs roughly 3-6 ms, so every frame of the move took 12-15 ms of UI time;
off the art they return at once.
- canvas.rs: the canvas notes when its view (rect, zoom, centre,
rotation) last changed. Until it has been still for HOVER_STILL_SECS
(0.15 s) it draws no hover highlight and keeps the cursor the tool last
gave for the same tool and modifier keys instead of asking again; it
asks for a repaint then, so both come back without the pointer moving.
Pointer moves still reach the tool as before.
- test: a_moving_view_leaves_the_art_under_the_pointer_unhighlighted_until_it_stops
(wheel frames draw no hover outline and keep the cursor; once still,
the outline returns and the cursor follows the art). It fails on main
at the first wheel frame.
Measured on an Apple M4 (release build, 1440x900 window, the PhotoCraft
app icon SVG fitted at 138%, Selection tool, Smart Guides and Snap to
Point on), median UI frame time over 60 wheel-pan frames with the pointer
resting:
before: over the art 12.3-15.0 ms, off the art 1.0 ms
after: over the art 0.4-1.0 ms, off the art 0.6 ms
The render readout is the art's own raster time on the render worker,
which this change doesn't touch.
The panels' scroll areas used egui's default floating scroll bar, which
is drawn over the right edge of the content: in a window too short for
the Properties panel it covered the Image Trace button's dropdown arrow
and the W/H constrain link, and in a dock column too tall for the window
it covered the panels' menu buttons and right-hand controls. Menus
(#2086, list_scroll_style) and the toolbar and dock icon column (#808,
strip_scroll) had already been given bars that keep clear of their
items; the panel bodies hadn't.
- widgets: panel_scroll shows a panel's scroll area with the menus' slim
solid bar (list_scroll_style), in a column of its own at the right
edge; scroll areas inside the body keep their own style.
- dock.rs: the docked Properties panel and Properties in a flyout,
floating group or shared dock group use it.
- floating.rs: a dock column too tall for the window uses it.
- panel_docking.rs: a panel's body in the shared dock uses it, laid out
as wide as the scroll area leaves beside the bar; the shared dock's
icon column scrolls with strip_scroll, as the dock's icon column does.
- test: the_properties_scroll_bar_covers_no_control (a 1440x600 frame of
the dock with an image selected: no control's rect meets the
Properties scroll bar). It fails on main with "Tracing Presets" (the
Image Trace arrow) and "Constrain Width and Height Proportions" under
the bar.
Verified live at 1440x600 with an image selected: before, the hovered
bar sat over the Image Trace arrow and the right ends of the Quick
Action buttons, and a dock column's bar over the Swatches panel menu;
after, both bars sit in their own 8 pt column and nothing is covered.
With the Artboard tool the arrow keys did nothing to the artboards: the
tool claimed only Escape, Delete and Backspace (tools/src/xform/artboard.rs
claims_key), so the arrows fell through to object.nudge in
ui-egui/src/shortcuts.rs, which moved the selected art on its own (or
nothing, with no art selected) and left the artboard where it was.
- artboard.rs: the tool claims the arrow keys and turns each press into
one artboard.move of the selected artboards that aren't locked, by the
keyboard increment (Preferences > General > Keyboard Increment, read
from the tool context as Touch Type does), Shift ten times as far as
object.nudge does, with moveArt from the tool's Move/Copy Artwork with
Artboard option, as a drag sends it. One undo step per press; nothing
during a drag or with only locked artboards selected.
- tests: arrow_keys_nudge_the_selected_artboards (tools: claims, step,
Shift, moveArt, several boards, locked, mid-drag) and
artboard_tool_arrow_keys_move_the_selected_artboards_with_their_art
(engine: the artboards and the art on them move, one step per press).
Verified live over --control: a 400x300 artboard with a selected 40 pt
square, Artboard tool, ArrowRight: before, the artboard stayed at x 0 and
the square moved to x 51 ("Move"); after, the artboard goes to x 1 and the
square with it ("Move Artboard"); Shift+Down moves both 10 pt; with
Move Artwork off only the artboard moves; increment 5 moves 5 pt.
The drag case in the report: a drag already moves the art, but only the
art that lies wholly inside the artboard and is unlocked and visible
(doc/src/lib.rs art_on_artboard; locked and hidden art only with
Move Locked and Hidden Artwork with Artboard). Art overlapping the
artboard's edge stays behind by design, which may be what the reporter
saw when dragging.
- ui-egui: the font-install poll (#2576) only runs on native builds, but its
timestamp field, its initializer and FONT_INSTALL_POLL were compiled for
wasm too, where nothing read them. They're gated like the code that reads
them.
- engine: `Session::store_profile` is only called by the native Load
Profile (cmd/colormgmt.rs); gated the same way.
- app (macOS): `App::logic_frame` checked `!file_dialog_open()` twice, one
`if` nested in an identical one, with the comment repeated. One check.
`cargo xtask wasm` runs `cargo check` without `-D warnings`, so these
warnings didn't fail CI; after this change it prints none.
With image mips on (the canvas always sets them), drawing an image
fetched its full decoded pixels first, decoding them again when the
shared image store no longer held them, and only then looked up the mip
it draws. The full pixels are only peeked, so they are the least
recently used entries and the first the store drops over its budget
(1 GiB on macOS and Windows). Once a document's full-size images exceed
it (20 photos of 6000 x 4000 px are 1.9 GB as RGBA), every frame decoded
every image again, though only their small mips were drawn: moving one
image redrew seconds later.
The store now remembers the size each blob's pixels decode to, so the
mip level is found without the pixels, and a mip the store holds is
drawn (and marked used) without fetching or decoding them. Without the
mip, at full size, or with mips off (exports) drawing is unchanged.
Measured on an Apple M4 (load average 15-24 from other work), 20 noise
PNGs of 6000 x 4000 px fitted in a 2880 x 1800 view, one moved 5 pt per
frame: before, 20 decodes and 1.1-3.0 s every frame; after, the first
frame decodes 20 and the later ones none, in 3-5 ms.
The readout was a moving average of 1/dt over any two consecutive frames.
An idle window wakes now and then (a timer, a control request, a repaint
after input) and each wake-up draws two frames about 4 ms apart, the second
one not waiting for vsync because the swapchain has a free image. Every
wake-up pushed the average up by 0.1 x 240 and the idle gap pulled it down
by only 0.1 x 2, so at rest it climbed past the display's refresh rate
(105 fps on a 30 Hz monitor that was drawing 4 frames a second). As soon as
the pointer moved and frames became vsync-paced, it fell back to the real
rate, which looked like a slowdown.
- frame_rate: frames counted over the last second of continuous drawing.
A gap more than 100 ms longer than the previous frame's own work means
the window was idle and starts a new count. The readout changes only
after half a second of drawing, so a wake-up's frames keep the last
rate. Slow frames still read slow, because their gap is their own work.
- chrome: "– fps" until the window has drawn continuously, instead of 0.
Measured on an Apple M4 driving a 30 Hz 4K display: drawing continuously
reads 30; at rest with a wake-up every 0.5 s it stays 30 (before: 45 -> 105).
In a clip group or a clipping layer the bottom child is the clipping path.
Object › Arrange moved the selected children with no regard for that slot:
Send to Back (and Send Backward from just above it) put the art under the
clipping path, so the art became the clipping path, lost its role as art,
and the group drew nothing. Bring Forward / to Front on the clipping path
itself would likewise have made art the clip. Layers-panel moves already
keep the clipping path at the bottom (layer.rs, `index.max(1)`); arrange
didn't.
- object.rs `arrange`: in a parent that clips, the clipping path is left
where it is and the art is arranged within the slots above it (Back goes
to just above the clipping path, Backward stops there).
- tests: Back, Backward, several objects, and Forward/Front on the
clipping path in a clip group (with a render check); Send to Back in a
clipping layer.
Taha-Mahmoodi