Envelope Options stored Anti-Alias and Preserve Shape Using, but nothing that
draws an image read them: a distorted image was always smoothed and always
cut into clip groups.
- With Anti-Alias off, each piece of a distorted image is marked pixelated
(saved as `pixelated`). The renderer draws it with nearest-neighbor
sampling and without mips, SVG export writes `image-rendering: pixelated`,
and PDF export writes it without /Interpolate.
- With Preserve Shape Using Transparency, each piece is the image with a
white opacity mask in the piece's shape instead of a clip group. The
renderer paints an image whose mask is one plain white path straight into
that path, as it does for an image in a clip group, when the path lies
inside the image's frame.
Docs: docs/mcp.md, docs/target-app-parity.md.
Edits refuse to leave anything beyond MAX_COORD (doc_sane checks the
artboards, the guides, the ruler origin and the objects an edit touches),
but the native loader didn't check them. A file with one object, artboard,
guide or ruler origin out there opened without a warning, and then every
move failed with "result would exceed the canvas", in-range objects
included; saving kept it that way.
Document::drop_beyond_canvas removes objects whose geometric bounds aren't
finite or reach past MAX_COORD (a layer or non-clipping group keeps its
children within it), artboards with their artboard guides, and guides out
there, and puts the ruler origin back at the origin. load_file runs it
after the depth limit and reports what it did with the other load notes; a
file whose artboards all lie beyond the canvas is refused.
The reference app disables its Lock, Hide and Show All menus in
isolation mode. In VectorCraft every item of Object › Lock and
Object › Hide, Unlock All and Show All worked there.
In isolation mode, object.lock, object.lock.above,
object.lock.otherLayers, object.hide, object.hide.above,
object.hide.otherLayers, object.unlockAll and object.showAll are now
off. Their menu items are greyed out, and a call returns an error
that ends "exit isolation mode first". The Contextual Task Bar's
Lock button is greyed out too, and the Properties panel's Quick
Actions leave out Lock. Pattern editing mode and opacity mask editing do not count as
isolation mode for these commands. The check is outside_isolation in
cmd.rs, with has_selection_outside_isolation beside it.
Docs: docs/mcp.md, docs/target-app-parity.md.
A closed arc from Arc Segment Tool Options, an Arc tool drag or
shape.arc with lengthX and lengthY closed through the box's bottom-left
corner with Base Along X, which makes a convex arc a quarter-ellipse
sector. With Base Along Y it closed through the corner a convex arc
bulges toward, so it drew the thin sliver between the arc and that
corner instead.
Such a closed arc now closes through the corner across from the one a
convex arc bulges toward with either base, so the Y-axis arc, open or
closed, convex or concave, is the X-axis arc mirrored top to bottom.
Open arcs and arcs with Base Along X are unchanged. The two-point
shape.arc form, whose params list no baseAlong, now closes through
(x1, y2) whatever baseAlong says.
Docs: docs/target-app-parity.md.
The merge test covered adjacent matches and a path of another color
between two matches. It now also covers a matching path the stroke
does not touch, between two it touches: both touched paths merge into
the upper one and the untouched path is left as it was.
A Blob Brush stroke used to merge with every closed path of its fill
color and no stroke that it touched, anywhere in the document, across
layers and groups and past objects of other kinds. It now merges with
the topmost such path it touches, and with the other touched matching
paths in the same group or layer that are next to that path in
stacking order, with no other kind of object between them. Merge only
with selection still limits the candidates to selected paths. The
merged path keeps the topmost path's id and place.
The reference app places a new layer below the selected layer with
Ctrl+Alt-click (Command+Option-click) on the New Layer button, and a
new sublayer at the bottom of the layer list with the same click on
the New Sublayer button. In VectorCraft that click opened Layer
Options, and New Layer and New Sublayer always added the layer above
the current layer and the sublayer on top of the layer's contents.
layer.new takes below, which adds the layer directly below the current
layer at its level, and layer.newSublayer takes bottom, which adds the
sublayer under the parent's contents. In a clipping layer the new
sublayer goes above the clipping path. Each is one undo step, and
below is refused together with top. A Cmd/Ctrl+Alt-click on the
Layers panel's New Layer and New Sublayer buttons runs them. A plain
click, an Alt-click and a Cmd/Ctrl-click work as before.
Docs: docs/mcp.md, docs/target-app-parity.md.
A .ai file whose editing data has no layers, as swatch libraries are
saved, opens from its PDF part, and the PDF part has the default
swatches. layered_ai applied the editing data's palette only when it
built the document from the layers, so Other Library… listed the
default swatches for such a library, and File › Open gave the file the
default swatches as well. A .ai whose layers can't be used (an image
too large for the editing data's reader, or a read error) lost its
palette the same way, and so did an EPS that opens as its page.
The reader now reads the palette on its own (ai::palette, which shares
palette_swatches with the layer reader), and its swatches and color
groups replace the default swatches in those three cases. Spot and
pattern swatches of the PDF part's art that the palette does not have
follow the palette's swatches.
Fixes#4100. Part of #3349.
The Type tool heard of Undo and Redo only when they ran as top-level
commands. As steps of command.batch, they left the caret where it was,
and the tool took the batch for an outside edit. Session::execute now
tells the active tool about each nested edit.undo and edit.redo.
command.batch also tells the tool when it begins, ends or rolls back.
The Type tool keeps its caret and its edit records at the start of each
batch and puts them back when a failed batch rolls the documents back.
While the Type tool edits type, Edit > Undo and Redo (the menu, the
keyboard, the History panel and MCP all run edit.undo and edit.redo)
left the caret at its byte offset, clamped to the text. Typing "abc"
before "XYZ" and pressing Undo left the caret after "XYZ".
The engine now tells the active tool about edit.undo and edit.redo
before it finishes the command. The Type tool keeps the edited text and
a list of the steps that changed it, and puts the caret, with nothing
selected, at the end of the span the step changed: where typed text
began after its Undo, after restored text after the Undo of a deletion,
and after retyped text after Redo. For the steps the tool made, it keeps
where the span ended, so text that repeats the characters next to it
puts the caret at the right end. A step that changes no characters,
such as a formatting change, leaves the caret and the selection as they
are. Typed lists keep their caret positions.
GIF, WebP and Targa files hold no resolution field, so document.formats and
the export docs now say that ppi, and the resolution that width or height
gives, are stored in the file except in those three formats.
PNG, PNG-8, JPEG, WebP, GIF, TIFF, BMP, Targa and PSD exports take
`width` or `height` in pixels. Each file is scaled so that its own
artboard, `rect` or art bounds is that wide or high, so with
`useArtboards` every file gets its own scale. `width` wins over
`height`, and either wins over `ppi` and `scale`, in the order an
Export for Screens row reads its size. The scale stays within 0.01 to
64 pixels per point and the file stores the resolution that results.
A value that is not a positive number is refused.
The per-file scale comes from `ScreenSize::scale_for`, which Export for
Screens rows already use. Rows keep their own size: their format
settings still drop `width` and `height`. `document.formats` lists the
two options for every raster format.
The check after the last file returned the cancelled error for every
export. A folder export without a ZIP has written each file by then, so
a cancel that came during a final step that writes nothing (a row whose
file repeats the one before it) reported a finished export as
cancelled. The check now runs only for exports that still hold their
files: those without a folder and those that make a ZIP.
A cancel that came after the last file's check was not seen: an export
without a folder returned its files or its archive anyway. The export now
checks once more after the last file and again after deflating the
archive, so a cancel returns the cancelled error and nothing is written
or returned.
zip::deflate takes a `stop` callback that it calls before each 1 MiB of
a file. Export for Screens passes the cancel state, so deflating stops
at the next MiB. file.package passes one that never stops, so its
behavior does not change.
With `zip: true`, document.exportForScreens deflated nothing: every file
went into the archive stored. Each file is now deflated, or stored when
deflating does not make it smaller, as file.package's archive does. The
file list, the names and the reply's fields stay the same; `bytes` and
the archive's contents change. Entries keep the fixed 1980-01-01 date, so
the same files still give the same archive.
The deflating writer moves from package.rs into fileio/zip.rs as
`zip::deflate`, beside `zip::store`, and file.package calls it there with
its own date and error text. The test reader `zip::entries` reads
deflated entries too. The `zip` option's text in docs/mcp.md,
docs/control-protocol.md and the command's params says the files are
deflated.
An Alt-click right of a short layer name, inside the wider rename
target, no longer replaces the selection. The isolated layer's own row
now selects the art inside it (layer.selectAll), where it cleared the
selection before. The doc line says the click selects visible, unlocked
art and clears the selection on a hidden or locked layer.
A click with the Scissors tool cut the path whether or not Alt was
held, and the cursor stayed the Scissors cursor. The reference app's
shortcut list has "Switch Scissors tool to Add Anchor Point tool" on
Alt (Option).
While Alt is held, a click with the Scissors tool now does what a click
with the Add Anchor Point tool does: on a segment it adds an anchor
point with path.insertAnchor in one undo step, and the path is not cut.
The cursor is the Add Anchor Point tool's while Alt is held. When Alt
is let go, the cursor is the Scissors again and a click cuts.
Docs: docs/target-app-parity.md, crates/tools/src/cursors.toml.
The reference app's shortcut list switches between its Type tools with
Shift and with Alt (Option). VectorCraft's Type tools had no such switch.
While no type is being edited, a press with Shift held makes new type
with the chosen tool's horizontal or vertical counterpart: Type and
Vertical Type, Area Type and Vertical Area Type, Type on a Path and
Vertical Type on a Path. A press with Alt held makes it with Type on a
Path in place of Area Type, or the reverse, in the same writing
direction. With both keys held, both switch. The keys held when the
press starts are the ones that count. The pointer shows the tool the
press acts as, and the chosen tool stays selected. While type is being
edited, Shift and Alt work as they did.
Docs: docs/target-app-parity.md.
The reference app switches the units of measurement with
Ctrl+Alt+Shift+U on Windows and Option+Shift+Command+U on macOS.
VectorCraft could set the document's units with document.setUnits or
a right-click on a ruler, but no command stepped through them and no
key did it.
document.switchUnits (Switch Units, default Cmd+Alt+Shift+U) sets the
document's units to the next one in the order a right-click on a
ruler lists them: Points, Picas, Inches, Millimeters, Centimeters,
Pixels, Feet & Inches, Meters, Yards, Feet, and from Feet back to
Points. Each press is one undo step, and the reply names the new
units. Keyboard Shortcuts lists it under Document.
The precise cursors shortcut-editor test used Cmd+Alt+Shift+U as a
free key; it now uses Cmd+Alt+Shift+Q.
Docs: docs/mcp.md, docs/ui-parity.md, docs/target-app-parity.md.
Deleting a point of a finished path with the Curvature tool sent
path.removeAnchor, which refits the two segments that met at the point so
one curve follows them. The path kept the bend of the deleted point
instead of curving through the points that remain.
The tool now sends path.removeAnchor with curvature: true, which uses the
new curvature_remove. Each smooth neighbor takes new handles from its new
neighbors, and an end beside one of them takes its end handle again. A
corner neighbor keeps its handle directions, and its handle on the new
segment scales with the segment's chord. With two points left, the points
that curve lose their handles. Object > Path > Remove Anchor Points and
the Delete Anchor Point tool still refit the curve.
An anchor added to a path that wasn't selected now selects that path,
the new anchor direct-selected, so the point shows and can be edited at
once instead of only after a Direct Selection click. Undo restores the
old selection; a path already selected (the Pen's Auto Add) keeps it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Offset Path outlined an open path on both sides, so a shape whose last
anchor sits on its first came out as a ring with a hole. Such a path is
now offset as a closed shape: its last anchor merges into its first,
which takes the last anchor's in-handle, with the same 1e-6 test that
Object › Path › Join uses for one path (crates/engine/src/cmd/path.rs).
The Offset Path effect calls the same function.
edit.colors.saturate's params text names intensity as required, but a
call without it, or with a string such as "40", read the intensity as
0 and still ran Saturate. It changed no visible color but recorded a
Saturate undo step.
edit.colors.saturate now reads intensity with f64_req, as shape.ellipse
reads its box. A missing, null or non-number intensity is the error
"missing number `intensity`", which changes nothing and records no undo
step. A number outside -100..100 still counts as its nearest end, and
the params text says so.
Saturate returns a color whose saturation stays the same, as at an
intensity of 0, as it was. The HSB round trip rewrote most such colors
in their last bit, so they counted as changed and lost their swatch
link.
In the same file, edit.colors.adjustBalance read a channel that is not
a number as 0, and color.harmony read steps or amount that is not a
number as its default. Both now refuse it with check_numbers, as the
commands docs/mcp.md lists for #1952 do. A null is still read as a key
left out.
Docs: docs/mcp.md, docs/target-app-parity.md.
layer.new, layer.selectAll, layer.target, layer.move, layer.template,
layer.pasteRemembersLayers and layer.releaseToLayers read their on/off
params with bool_or or as_bool, which falls back to a default for a
value that is not true or false. layer.new {"top": "true"} added the
layer above the current one instead of on top, layer.template
{"on": "false"} and layer.pasteRemembersLayers {"on": "false"} turned
the option on when it was off, layer.move {"copy": "true"} moved the
rows instead of copying them, layer.selectAll and layer.target with
"add": "true" replaced the selection, and layer.releaseToLayers
{"build": "true"} released the objects without building up copies.
These commands now fail with "`<key>` must be true or false" for a
value that is not true, false or null, as the artboard and Object menu
commands do, and leave the document, the selection, the highlighted
rows and the undo history as they were. layer.newSublayer, which takes
layer.new's options, refuses such a top too. A null is read as a key
left out.
docs/mcp.md also lists showCenterMark, showCrossHairs and
showVideoSafeAreas among the keys artboard.new and artboard.setProps
refuse in this way, as they already did.
Docs: docs/mcp.md, docs/target-app-parity.md.
artboard.setProps refused an x, y, width or height given as text, but
artboard.new read such a value as the key left out: {"x": "10"} added
an artboard at the default position and recorded an undo step. Both
commands read a name that is not a string, such as 5, as no name, so
artboard.new gave the default name and artboard.setProps kept the old
one.
artboard.new now refuses an x, y, width or height that is not a number
with "`<key>` must be a number, not <value>", the message
artboard.setProps gives, and both commands refuse such a name with
"`name` must be a string", for each entry of artboard.setProps'
artboards too. They refuse it before they change anything. A null is
still read as the key left out.
Docs: docs/mcp.md, docs/target-app-parity.md.
Found while testing crates/engine/src/cmd/create.rs:584 (dividers)
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Found while testing crates/engine/src/cmd/place/rtf.rs:111
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Writer::fresh_id probed prefix-1, prefix-2, ... from 1 on every call,
formatting and hashing each candidate, so writing N gradients, clip
paths or masks cost O(N^2). Ids are never released, so every number
below the last one handed out for a prefix is taken: remember where
the search for each prefix stopped and resume there. The ids written
are the same as before.
examples/still-life.vectorcraft (12k gradients, 6.3 MB SVG) exported
in about 8.6 s (median); it now takes about 0.15 s. Output is byte-identical
for every document in examples/.
Add crates/engine/examples/export_bench.rs to time document.export
on the example documents, and a test pinning the generated id
sequence when object names already take some of the numbers.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review follow-ups:
- items_per_target maps each requested row on its own. Rows deleted or
duplicated together, one shared and one not, fell back to the literal
indices in every object, so the other objects could lose the wrong
items. The first object takes its own rows; each other object takes
its copies of the shared ones only.
- effect.dialog takes `aligned` (the panel passes it on a mixed
selection's shared effect row), and its OK and preview run
effect.setParams with it, so the dialog edits each object's own copy
of the effect instead of the first object's index in every object.
Tests: removing a shared stroke with an unshared fill leaves the other
object's fills; an aligned dialog's command carries `aligned`.
With several objects selected whose appearances differ in anything (two
circles whose fills differ but whose stroke and effect are the same),
the Appearance panel showed only "Mixed Appearances": no stroke, effect
or opacity row, so attributes the objects share couldn't be edited
together. It now lists what they share, as the issue describes:
- `vectorcraft_doc::appearance::SharedStack` (with `align`, a stacking-
order longest-common-subsequence match) gives the fills, strokes and
object-level effects every object has an equal copy of, and where each
is in every object's own stack.
- The panel's object row names the objects' common kind ("Path",
"Ellipse"; "Mixed Objects" only when kinds differ, as the Control bar
does), with an italic "Mixed Appearances" row under it when they
differ, then only the shared rows with their usual controls. The
Opacity row reads "Mixed" when opacity or blend mode differ (an edit
sets them all). The object row shows no fill chip when the fills
differ. Rows aren't dragged in a mixed selection: where a row would go
in each of several different stacks isn't one place.
- The engine maps a shared row to each object's own item: the active
item (a row of the first selected object's stack) always does, so the
Color, Stroke and Transparency panels edit every object's matching
item, and the item and effect commands do with `aligned: true`
(appearance.setItem / removeItem / duplicateItem, effect.remove /
duplicate / setParams / lightingPreset, transparency.set), each as
one undo step. moveItem and effect.move refuse aligned rows of
differing stacks. Without `aligned`, explicit indices keep their
meaning for single objects and scripts.
- "Mixed" is added to every catalog.
Tests: the shared stack of the issue's case, of a stroke at another
index, of nothing alike and of equal stacks, and that matches never
cross; editing the shared stroke edits both strokes (at index 1 and 2)
in one undo step that restores both, also through stroke.set on the
active row; removing a shared effect and adding a fill reach both, and
reordering differing stacks is refused; the panel lists kind, note,
Stroke, Feather and Opacity, not Fill, the full stack for alike objects
and "Mixed Objects" for a path and type. The existing object-row test
now expects the kind for two rectangles and the shared rows when their
opacity differs.
dotCooCoo