A Shift-click on a guide adds it to the selected art (guide.select
toggle keeps the art), and Shift-adding or -removing art keeps the
selected guides; a new selection still clears them. With both selected,
Delete deletes the art and the guides in one undo step (not when ids are
given).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CC5DiGJEZ5ADaQMBEQk4W1
A PDF that names a font without embedding it (here a standard Helvetica)
is drawn with a stand-in, so comparing its glyphs with the installed
family's styles picks whichever style happens to look most like the
stand-in: with a large Helvetica family installed, Helvetica became
Helvetica CE Medium (tests_import_fidelity::text_becomes_point_type
failed on such a Mac). Fonts without a font program now keep the face
their name resolves to; embedded fonts are still compared as before.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CC5DiGJEZ5ADaQMBEQk4W1
The link watch wakes the app every two seconds only while the last look
found links (068a934, #967). A file placed less than two seconds after a
look that found none asked for no more frames, so while the window sat
idle nothing stamped it; a change made meanwhile was taken for the file
as first seen and never acted on. The watch now remembers the active
document's uid and revision at each look and, when they differ on a
later frame, asks for one wake when the next look is due. There is
still no per-frame walk of the links, and no two-second wakes for a
document without links.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
$INSUNITS 22, 23 and 24 (US survey inch, yard and mile, as Autodesk's INSUNITS reference
lists them) had no size or name, so the drawing opened as unitless (millimetres, or inches
with $MEASUREMENT 0) and document.dxfInfo said Unitless.
They now derive from the US survey foot (1200/3937 m) like code 21, with their names.
The test reads each survey code's unit name and default scale and opens a survey-inch line at
7200.0144 pt.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Conflicts:
# ROADMAP.md
A root such as width="96pt" height="96" took CSS pixels (0.75 pt) along x and points along y,
so a square came in 24 x 32 pt and the artboard 96 x 96 instead of 96 x 72; with a viewBox the
art was stretched too.
When either root side is absolute, both axes now take CSS pixels at 96 per inch (usvg already
parses at that resolution): the unitless or px side is the same CSS length it would be in a
browser. The root scale is one number.
The test opens width="96pt" with height 96, 96px and 72pt, and width 128 with height 72pt: a
96 x 72 artboard and a 24 x 24 square; with viewBox 0 0 32 32 a 72 x 72 square at x 12 (meet),
and with preserveAspectRatio="none" the whole artboard.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Conflicts:
# ROADMAP.md
Cursor::precise listed every Pen state but PenJoin (added later, #794), so with Use Precise
Cursors on the Pen still showed its join badge over another open path's end.
PenJoin is a crosshair too under the preference.
The Pen join test now also checks the pointer over the other path's end is the crosshair with
the preference on (and PenJoin with it off, as before).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Conflicts:
# ROADMAP.md
shapes::polar_grid drew at least one radial divider (`radial.max(1)`), so radial 0, which
the command and the Polar Grid tool (Left arrow while dragging) accept, still drew a spoke.
Zero now draws none; the loop never divides by it.
The test checks 0, 1 and 2 radial dividers with no concentric ones give 1, 2 and 3 paths.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Conflicts:
# ROADMAP.md
Reversal read the midpoints after Gradient::reverse had reordered the stops, so each stop
took the wrong segment's midpoint: red-to-blue with its midpoint at 20% reversed to a 50%
midpoint instead of 80%, moving the transition.
Gradient::reverse now mirrors each segment's midpoint with it, so a reversed gradient is the
same ramp the other way round; paint.editGradient's own (wrong) pass is gone, and the SVG
importer's reflect spread, which had a correct copy of the same mirroring, uses it too.
Tests: two- and three-stop gradients with off-centre midpoints sample the same reversed (t vs
1 - t) and come back after two reversals; paint.editGradient reverse turns a 0.2 midpoint into
0.8.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Conflicts:
# ROADMAP.md
The menu item sends no artboard, so the engine fell back to the first one even with another
active in the Artboards panel.
The frontend fills in the active artboard (the view's, as the Artboards panel and navigator
show it) for select.allOnArtboard, as it already did for Paste in Place / in Front / in Back
(#693); the two now share one block, and the index is kept within the artboards. An explicit
artboard still wins; the engine is unchanged.
The test makes artboard 2 active, then 1, and invokes the menu item each time, then passes
artboard 1 explicitly with artboard 1 active.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Conflicts:
# ROADMAP.md
The desktop app's ui.render (MCP `screenshot` when connected) always rendered the first
artboard and ignored `artboard`, which the headless backend honours.
It now renders the 0-based `artboard` (the first without one, an error for one that doesn't
exist), as headless does. control-protocol.md and the control.rs list show the parameter.
The test drives ui.render through the control channel on a 64 x 64 and a 96 x 48 artboard:
no index and 0 give 64 x 64, 1 gives 96 x 48 (48 x 24 at scale 0.5), 2 is an error.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Conflicts:
# ROADMAP.md
add_text passed `color` to text.create only, and text.createInPath built its own character
style without a colour, so type on a path and path-based area type took the current fill.
text.createInPath now uses the same new-type style as text.create (size, font, style, color,
else the current fill), and add_text sends `color` on both routes.
The MCP test draws a blue rectangle (the current fill), turns it into on-path and area type
with color #ff0000, and reads red back through inspect_document.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Conflicts:
# ROADMAP.md
With every layer hidden or locked, the insertion layer fell back to the top layer anyway, so
drawing a shape, typing, pasting, making a graph, a flare, a grid, trim marks or a symbol put
the art on a hidden or locked layer: invisible, or unselectable by Select All. Illustrator
refuses to draw on a locked or hidden target layer (public forum answers, e.g.
https://community.adobe.com/t5/illustrator/can-t-draw-a-new-object-in-my-document/m-p/11107182).
DocState::target_parent is insertion_parent for new art, failing with "the target layer is
locked or hidden" when it would be a layer that isn't editable; those creation paths use it.
Another layer that is shown and unlocked still takes the art, as before. The message is in the
de, es, fr, it, ja, ru and uk catalogs. File > Place is left as it is (another branch reworks
it).
The test hides, then locks, the only layer: an ellipse, a polar grid, text and a paste each
fail with the message and leave the document unchanged; a new layer then takes a rectangle.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Conflicts:
# ROADMAP.md
Character and paragraph styles checked a new attribute against the keys its base style
serializes, and the JSON leaves out Auto leading and kerning, no mojikumi and other defaults.
So charStyle.new / setAttrs refused leading or kerning, paraStyle.new refused mojikumi, and a
style captured from text with explicit leading failed to apply to Auto-leading text, to
duplicate or to come back out of a library.
The check now also accepts every field name the style type deserializes (read from serde's
derive), still refusing unknown keys and the file-only justify_auto.
The test creates and edits styles with kerning, leading, small caps, mojikumi and direction,
applies a captured leading style to fresh text with and without Clear Overrides, duplicates it,
uses it from a library, and checks `nope` and `justify_auto` are still refused.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Conflicts:
# ROADMAP.md
Asset Export, File › Export, Save for Web and Object › Rasterize drew
each shape with its own smoothed edge, so where two shapes met edge to
edge (Divide, Shape Builder, abutting art) both half-covered the pixels
on the line and the background showed through as a thin ghost border.
With Art Optimized (the default) these now supersample, as the option's
name in Illustrator says: the art is drawn with hard edges at four times
the size (three or two times for large images, within a 40-megapixel
budget and the renderer's 65,535-pixel side) and each block averaged
into one pixel. Pixels shared by two shapes are fully owned by one or
the other at the finer size, so no seam is left, and the outer edges
are still smooth. The canvas keeps shape-by-shape smoothing for speed;
None and Type Optimized are unchanged.
RenderOptions::supersample turns it on (exports and Rasterize set it).
A render test checks two triangles sharing a diagonal: a seam with
shape-by-shape smoothing, none supersampled, and the factor's limits.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The headless test taken from #1015 set `RawInput::modifiers`, which egui
0.36 doesn't have, and clicked left of the spinner (the right-aligned
label's text ends before its box). Modifiers now come as
`ModifiersChanged` events, as the Appearance panel tests send them, and
the click is placed from the field's text. It also checks Ctrl/Cmd steps
a tenth. It fails with the spinner ignoring modifiers (15 pt for 24).
rustfmt.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The PR kept a field's integer kind only while its value had been an
integer: typing 2.5 in Columns and then 3 sent 3.0, and the mosaic fell
back to 10 × 10 again. A whole number typed in any command dialog is now
an integer (`typed_number`) whatever the field held before; commands
that take any number read integers as well, and a fraction stays one.
Test: Object › Create Object Mosaic… opened from its menu item, 3 typed
over Columns and Enter makes a 3 × 10 mosaic (it made 10 × 10 before,
#1000).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Same checks in a third of the code; the corners destructure into the
four points instead of being indexed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The PR added `checked_anchor_from_json` next to the lenient
`anchor_from_json` and dropped path.rs's import of the latter, which
`path.appendAnchor` still called, so the engine didn't compile.
`anchor_from_json` is now the strict parser itself (handles through
`finite_numbers`), used by path.create, path.setAnchors and
path.appendAnchor alike: the Pen's appended anchor with a malformed
handle fails too instead of losing the handle. Tests: a non-boolean
`closed` and a malformed appended handle change nothing. rustfmt.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
`text.reshapeArea` read its anchors with the lenient `parse_refs`, so a
malformed reference among good ones moved only the good ones. It now
uses `checked_refs` as well, and `parse_refs` (no other user) goes.
Test: a mixed list fails without moving anything.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A rejected Stroke panel edit with nothing selected leaves the next
object's stroke as it was; an accepted one sets its weight and dashes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- `finite_numbers::<N>` (cmd/mod.rs) reads an array of exactly N finite
numbers, nothing dropped or padded; `rect` uses it, and the rect cut,
Free Distort and anchor handles reuse it in the PRs stacked on this one.
- `at` is parsed once (`at_param`) instead of point_param plus a second
shape check that indexed the params; `at: null` means no position, as
null does for the other optional params, and gets a test.
- rustfmt.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Eight raster effects on the pixel-filter pipeline
(vectorcraft_effects::pixel::brushstrokes), under a new Effect › Brush
Strokes submenu between Blur and Distort in the Raster Effects section:
- Accented Edges (edgeWidth pt 1..14, default 2; edgeBrightness 0..50,
38; smoothness 1..15, 5): the edges of the softened object painted
towards white chalk at high brightness or black ink at low brightness;
Smoothness also leaves the faint edges out. The shape stays as it was.
- Angled Strokes (directionBalance 0..100, 50; strokeLength pt 3..50, 15;
sharpness 0..10, 3): light areas in right diagonal strokes, dark areas
in left diagonal ones; 0 makes every stroke a left diagonal, 100 every
stroke a right diagonal. Sharpness brings detail and bristles back.
- Crosshatch (strokeLength pt 3..50, 9; sharpness 0..20, 6; strength
1..3, 1): the detail kept under pencil hatching along both diagonals,
Strength passes of it, the edges roughened.
- Dark Strokes (balance 0..10, 5; blackIntensity 0..10, 6;
whiteIntensity 0..10, 2): short strokes towards black in the dark
areas, long strokes towards white in the light areas.
- Ink Outlines (strokeLength pt 1..50, 4; darkIntensity 0..50, 20;
lightIntensity 0..50, 10): black outlines over the edges, ink hatching
in the shadows, shadows darkened and highlights lightened.
- Spatter (sprayRadius pt 0..25, 10; smoothness 1..15, 5): each pixel
shifted by noise up to 0.4 x the radius; Smoothness makes the grains
larger and smoother.
- Sprayed Strokes (strokeLength pt 0..20, 12; sprayRadius pt 0..25, 7;
strokeDirection Right Diagonal, Horizontal, Left Diagonal or Vertical):
strokes along the direction, scattered by up to 0.3 x the radius.
- Sumi-e (strokeWidth pt 3..15, 10; strokePressure 0..15, 2; contrast
0..40, 16): wet strokes, ink bleeding from the darks along paper
fibres, more ink with pressure and an S-curve contrast for rich blacks.
It works in value only: hues and saturation stay, and so does the
object's shape (no fringe in the transparency).
A stroke averages the colour along a segment through the pixel (16
samples at most, the art softened first when they lie more than a pixel
apart); strokes lie in rows around the object's centre, each row
starting at random and each stroke shifted along itself, with streaked
noise for the bristles. Everything is in document space, so the effects
look the same at any zoom or resolution, move with the object and draw
the same every time; outsets and reaches are exact. On a 2000 x 2000
raster in release the slowest settings take 0.3-2.0 s (Crosshatch at
50 pt with 3 passes the most).
Shared helpers instead of copies: edges::sobel (Glowing Edges now uses
it, unchanged), distort::displace and texture::Tones are pub(super).
Catalogue rows with parameter docs, menu entries, Sprayed Strokes'
direction dropdown, translations (de, es, fr, it, ja, ru, uk, zh-hans,
zh-hant; menu labels in cs and pt-br), tests (parameters and clamping,
outsets and reaches, each filter and option showing, chalk and ink on
the edges only, shadows and rich blacks, Sumi-e keeping the shape and
hues, hatching keeping the tone, strokes smearing along their direction,
Spatter's radius, resolution and move invariance, canvas, multithreaded,
PDF and SVG, the command sweep with junk parameters, the menu) and
ROADMAP.
Behaviour from public documentation only:
https://helpx.adobe.com/photoshop-elements/using/brush-stroke-filters.htmlhttps://www.computerhope.com/jargon/p/photoshop-sprayed-strokes.htm
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Revolve dialog and gizmo strings in German (the catalog landed after
this branch was written), with Kreiseln and Aussehen umwandeln as the
catalog already names them. Intensity, which main's Texture filters
added meanwhile, is kept once per catalog. README's missing list says
what 3D still lacks beyond the initial Revolve.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
As Window's other panel items do, and as the reference's Window menu has
it, Window › Variables is the panel itself instead of a submenu of
eighteen commands. The panel menu gains Make Text Dynamic, Make
Visibility Dynamic, Unbind Variable and Delete Variable beside the
bottom bar's buttons, which share one reading of what the selection
allows; every command stays in the command palette. The labels only
the submenu used leave the catalogs.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Dataset and variable names are the user's: the panel's Data Set dropdown and
the dialogs' name dropdowns show them as they are instead of translating
them; the dialog's Kind dropdown shows translated kind names.
- Choosing the empty option in a dialog dropdown picked the first name.
- The panel reads the art once per frame for every row's italics, the dialog
no longer clones the document's variables every frame, and slices are
matched instead of indexed.
- "No document" and "Define a variable first" go through tl!, and the dialog
headings Variable Options / Edit Data Set, Text and Visibility have rows in
every complete catalog; the unused "{name} · {kind} · {count}" row is gone.
- German rows for every Variables string (the catalog predates this branch).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- A visibility variable could hide an object but never show it again: dataset
application, capture and binding refused any object that wasn't visible
(`is_editable`), so a row that hid an object left it hidden for good and a
capture after hiding it in the Layers panel didn't record it. They now use
`Document::variable_target`, which ignores the object's own visibility
(locked art and art on hidden or locked layers are still skipped).
- Renaming a variable renamed the active dataset when the two shared a name,
leaving it pointing at a dataset that doesn't exist.
- `variable.delete` and `dataset.delete` without `names` act on the
highlighted variable and the active dataset, so Window > Variables > Delete
Variable / Delete Data Set no longer fail with "missing `names`".
- `ids` is checked like other commands' explicit targets (a non-list no longer
falls back to the selection, bad ids are errors), names are trimmed and
capped at 255 characters, and the bind checks are shared by Bind and Make
Dynamic.
- Make Dynamic's fallback name no longer becomes " 2" for unnamed art.
- `dataset.list` reads the art once for every row's drift check, and
`Variables::prune` looks names up in a set.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Select Bound Object ran variable.unbind on the selection, so choosing it silently
removed the bindings; Delete Variable and Delete Data Set deleted every row. Now
the panel keeps a highlighted variable row (a click highlights and selects what
it drives, as the reference panel does): Select Bound Object selects what the
highlighted variable drives via select.set, Delete Variable deletes only it, and
Delete Data Set deletes only the current set. Variable Options and Rename Data
Set open their dialogs preselected with the highlighted variable / active set
(the double-click used to invoke variable.rename with a bare name, which only
errored). Unbind enables only over a bound selection; Make Text Dynamic only for
exactly one type object.
The highlight is per-document session state with its own command
(variable.highlight, no undo step, like layer.highlight): rename carries it,
delete clears it when stale. New engine tests (18) and panel tests cover the
reviewer's regression case, verified to fail on the old wiring.
Opening the panel from the Window menu worked, but collapsing the dock did not
show it: `ICON_PANELS` only registers a panel (so `window.panel` finds it) while
`ICON_PANEL_GROUPS` is what the collapsed column actually draws, and `variables`
was in the first list only. Sixteen panels are registered without a group; for
the ones meant to live in the column, the entry in `ICON_PANELS` alone is not
enough.
`variables` gets a row of its own at the bottom of the column, where the
reference app keeps it. Both lists now say which is which, so the next panel
does not repeat this, and a test asserts a registered panel reaches the column
(it fails without the group entry) and that every grouped id names a registered
panel with an icon.
`floating.rs`'s icon test counted the column's icons, which only means
"one left" while the column is tall enough to show them all — past that, one
scrolling into view replaces the one that left. It runs on a window tall enough
for every icon now, and checks the count is the number of registered groups'
icons rather than a number that happens to pass.
Dragging a panel into the icon column still doesn't work, and is not something
this PR changes: tearing a panel off makes it float, for every panel. Docking it
back into the column would be a dock feature of its own.
cargo xtask ci: all 7 steps pass on Rust 1.99.
The engine side was called a solid base, but the panel could only list and
apply: variables and datasets were made through commands, which most people
won't find. The reference app's Variables workflow is now what the panel does,
and every button runs a `variable.*` / `dataset.*` command, so the panel and an
agent drive the document the same way.
New commands (all also in Window › Variables, the palette and over MCP):
- `variable.makeTextDynamic` / `variable.makeVisibilityDynamic`: define a
variable of that kind over the selection, named after what it drives, and
bind it — one step, as the panel's two buttons.
- `dataset.capture`: record what the bound objects hold right now as a new row
("Data Set 1"… ) and make it active. This, rather than typing values, is how
rows are made: change the art, then capture.
- `dataset.update`: write what the art holds now into the active row. Without
an active row it says so rather than picking one.
- `variable.rename` / `dataset.rename` (Variable Options…, Rename Data Set…):
a name is how a variable is referred to, so renaming carries its bindings and
its values in every row with it.
- `dataset.list` reports `matches` per row: whether the art still holds what
that row says.
`Document::captured_values` is the one place that reads a bound object's value,
so capture, update and the drift check can't disagree about what the art says.
Panel: the current data set across the top as a dropdown with ◀ ▶; variable
rows with their kind's icon and the name of the object they drive ("Multiple
objects" when it is more than one), a click selecting it and a double click
opening Variable Options; a dataset row in italics once an edit has moved the
art away from what it says; Make Text/Visibility Dynamic, New, Unbind and Delete
below, Capture and Update on a second line; and a (≡) menu with the same
commands for the ones that act on a row rather than the selection.
Checked on Rust 1.99 (the current stable), which is what the review asked for:
cargo xtask ci, all 7 steps pass.
`bindings` mapped an object id to a single variable name, so binding an object
to a second variable silently dropped the first — a binding lost with nothing
to show for it. The reference app binds the selected object to the selected
variable, one click at a time, and nowhere says a second binding replaces the
first (https://helpx.adobe.com/illustrator/desktop/automate-visualize-data/
automate-actions/work-with-variables.html), so this allows several instead.
`bindings` is now a set of variable names per object. `dataset.select` walks
them in the order the variables are defined, which keeps two text variables on
one object deterministic (the later one's value is what is left, and `applied`
counts both). `variable.bind` reports `bound` (added) and `already` (was there
already); `variable.unbind` takes an optional `variable` to drop one binding
instead of all of them.
This is done before the field ships rather than after: `variables` has never
been in a released file, so the shape can still change for free. Once a version
with it is out, the same change would need a format version bump and a
migration, since `crates/format` keeps every version from 1 loading.
A strict superset of what worked before: every binding a document had is still
there and still applied, in the same order.
cargo xtask ci: all 7 steps pass.
Review of #761 turned up three real defects and a feature that was visible
but not usable. All of it is fixed here, each defect with a regression test
that fails without the fix.
- A binding outlived the object it named. Deleting art left its entry in
`variables.bindings` forever: it counted as a skip in every dataset
application, `variable.list` counted it, and the saved file grew with every
delete. `Document::prune_variable_bindings` now lets them go after each
edit, as `prune_assets` does for Asset Export's.
- `Variables::prune` cleared `active_dataset` by comparing it against the
*variable* names being deleted. A dataset named like a variable lost its
active state when the variable went. The two are separate namespaces; the
block is gone (dataset.delete already did it right).
- `variable.delete` reported the number of names asked for, not the number
deleted, so deleting a name that was not there still said `deleted: 1`.
- `dataset.set` (new): a dataset's values could only be set when it was
created, so changing one meant deleting and rebuilding the row — losing its
place in the list and the active one. This replaces the row's values and
reports what it dropped.
- Window › Variables, and dialogs for the "…" commands. Every command had an
empty menu path and the command palette runs those with `{}`, so nothing
could be defined from the UI at all. The commands now carry their menu
path, the "…" ones open a Variables dialog (a name and a kind; a name and a
field per variable, its kind's editor), and `dataset.set` opens with the
row's own values.
- Locked art is neither bound nor rewritten by a dataset, and says so.
- `set_plain_text` told a node that existed but was not type that it was
missing.
- The panel had Document Info's icon, so the two were indistinguishable in
the collapsed column; it gets its own, plus New/Delete buttons, and its
rows say what their count is.
Format round trips now cover variables (the testkit fixture defines both
kinds, binds them and applies a row), and ROADMAP/README no longer list the
feature as missing.
cargo xtask ci: all 7 steps pass.
