Window › SVG Interactivity was a disabled menu item. VectorCraft could
not attach JavaScript to an object's events or link JavaScript files
for SVG export.
The SVG Interactivity panel lists 19 events, from onfocusin to
onscroll. Pick one, type its JavaScript and press Enter to add it to
the selected objects; the list shows their events. Remove Selected
Entry removes the picked event, and the panel menu's Clear Events
removes all of them. Link JavaScript Files… opens the JavaScript
Files dialog, which adds a path or URL to the document's list and
removes the selected one. Its Choose… button puts a picked file in
as a file:/// URL. The commands are svgInteractivity.info,
setEvent, removeEvent, clearEvents, addFile and removeFile, and
ui.javascriptFilesDialog. Each edit is one undo step. A document
links at most 64 files.
SVG export writes each event as an event attribute on the object's
element and each file as a <script xlink:href> element. With the SVG
Tiny 1.2 profile, export leaves the events out with a warning. SVG
import reads the listed event attributes and the script links back
as text. It never reads a linked file or runs a script, and it leaves
inline scripts and the links after the first 64 out with a warning.
The native format keeps both.
An unnamed <g> with data-* attributes of its own around a single
object now imports as a group. Before, import folded it into the
object, which then took the group's data-* or, when it had its own,
dropped the group's. An unnamed <g> with event attributes stays a
group in the same way. The SVG import fuzz tests now include event
attributes and <script> elements.
Docs: docs/mcp.md, docs/control-protocol.md, docs/target-app-parity.md,
docs/ui-parity.md.
An Alt-click (Option-click on macOS) on the Artboards panel's New
Artboard used to add an artboard at once, as a plain click does. It now
opens Artboard Options for the new artboard, filled in with the name and
size that artboard.new {} would give it, and adds nothing yet. OK adds
the artboard with the fields where artboard.new {} would place it (20 pt
to the right of the last artboard) in one undo step, and makes it the
active artboard; Cancel adds no artboard. A plain click and a row
dropped on the button work as before.
The new UI command ui.newArtboard ("New Artboard Options…") opens the
dialog. It is listed after ui.newLayer, and its label has a row in each
catalog.
Docs: docs/mcp.md, docs/target-app-parity.md.
The Contextual Task Bar showed only with the selection tools and
selected objects. With the Artboard tool it never showed.
With the Artboard tool, the bar now sits under the selected artboards
and offers Resize, which opens Artboard Options for the active
artboard; Duplicate, which copies each selected artboard in one undo
step and selects the copies; Rearrange All, with two or more
artboards, which opens Rearrange All Artboards; and Export, which
opens Export for Screens with the selected artboards checked. The
objects' Fill, Recolor Artwork and Lock buttons are not shown with
the Artboard tool. "Resize" has a row in every catalog.
Docs: docs/target-app-parity.md.
With the Artboard tool in use, the Control bar showed Move and Scale
Artwork with Artboard and then the object controls (Fill, Stroke,
Opacity and the rest). The reference app's help lists the Control
panel, with the Artboard tool, among the places to add, delete,
rename, resize and position artboards, set their orientation and
reference point, and open Artboard Options.
While the tool is in use, the Control bar now shows the active
artboard's controls in place of the object controls: Preset, Portrait
and Landscape, New Artboard, Delete Artboard, the Name field, Artboard
Options, Rearrange All Artboards, the Transform link, Move and Scale
Artwork with Artboard, and the reference point with X, Y, W and H and
the Constrain Width and Height Proportions link. Each control runs the
command its Properties or Transform panel counterpart runs, and the
reference point is the one the Transform panel uses. In an 800-point
window the controls up to the Transform link fit, the two artwork
options show as their boxes with their names in tooltips, and the link
opens the position and size in a popover, as it does for objects.
Docs: docs/target-app-parity.md.
Type › Optical Margin Alignment was a disabled menu item. The reference
app's Optical Margin Alignment hangs Roman punctuation and the edges of
letters such as W and A outside the margins of every paragraph of a
type object.
type.opticalMarginAlignment (Type › Optical Margin Alignment) turns it
on or off for the selected type in one undo step, and the menu item is
checked while the type has it on. In every left-to-right paragraph of
point and area type, quotation marks, hyphens, full stops and commas at
a line's start or end hang wholly outside the margin there. Asterisks,
tildes, ellipses, en and em dashes, colons and semicolons hang by half
their width, and A, V, W, Y, v, w and y hang by a tenth of theirs. A
punctuation mark followed by a quotation mark hangs with it.
Left-aligned lines hang at their left margin, right-aligned lines hang
at their right margin, and centered and justified lines hang at both.
Lines break where they did before, and the space that justification
adds between glyphs leaves out the gaps next to a hanging punctuation
mark.
The setting is off for new type and for documents from before it, and
the type object saves it only when it is on.
The Type menu's checks for Optical Margin Alignment, Type Orientation
and the Type on a Path effects now read the type inside a selected
group, as the commands do.
Docs: docs/mcp.md, docs/target-app-parity.md, docs/type-parity.md.
Object > Expand with Fill expanded gradient fills only. A pattern fill
stayed a live pattern, and object.expand.info did not report it, so the
dialog's Fill option was disabled for it. In the reference app, Expand
with Fill turns a pattern fill into the pattern's tiles, and the shape
becomes their clipping mask.
The Fill step now turns each visible pattern fill of a path or compound
path into a clip group shaped like the object, holding the tiles the
pattern lays over the object's box with fresh ids. The group isolates
the tiles' blending and takes the fill's opacity, blend mode and other
effects. Geometry effects of the fill or the object, such as Offset
Path, shape the group's outline and the box the tiles cover. A path in
the tile art with a random Transform effect is baked with its own id
first, and a path with a brush that varies at random (calligraphic or
scatter with Random variation, or bristle) is baked where the pattern
defines it first, so every tile matches the canvas. The object's other
fills and strokes stay on copies of it in paint order, and Expand stays
one undo step. Type with a pattern fill expands the same way once
Object outlines it, and object.expand.info reports pattern fills under
fill. It also reports gradient fills under a new gradient key, and the
dialog enables Expand Gradient To only when the selection has one. A
pattern fill stays as it is when it lays no tiles over the box, when a
group or type in its art has a random Transform effect, when the
pattern is scaled, rotated, reflected, stretched or sheared and its art
holds a symbol instance, and when the pattern is rotated, reflected,
stretched or sheared, or scaled while its art has geometry effects, and
its art has raster or 3D effects, or a geometry effect on a group or
type.
PatternDef gains instances_in_with, which places the tiles with a given
Scaling. Expand has stroke weights, effect distances and the tiles' own
pattern paints scale with the pattern's transform. When the pattern is
rotated, reflected, stretched or sheared, or scaled while its art has
geometry effects such as Roughen, the strokes and geometry effects of
the tiles' paths become filled art first (bake_appearance, and Outline
Stroke's conversion for brushed strokes), and their live shapes become
plain paths, in their opacity masks too. Tiles place their opacity
masks with them, unlinked ones included (place_art), which also fixes
unlinked masks in pattern tiles in PDF, EPS, EMF and WMF export and in
Distort Pattern Fills.
Faults in the shared code are fixed for every caller. assemble
dropped the effects of a sole part, so Expand lost a lone gradient
fill's own effects and Outline Stroke lost those of a stroke with a
gradient along or across it. Outline Stroke also dropped a brushed
stroke's raster effects from its brush art. Expand replaced an object
whose gradient fill could not be expanded with a plain copy; it now
leaves it as it was. Expand left the geometry effects of a gradient
fill, and with Object off those of the object, on the expanded art,
where they reshaped every strip; they now shape the clip group's
outline, and the gradient fits it, as for pattern fills.
bake_appearance dropped a stroke's raster effects from its filled
outline and filled the outline of a stroke with a gradient along or
across it with a gradient across the outline's box; it now keeps the
effects and makes the gradient meshes clipped to the outline, as
Outline Stroke does. Envelope Options > Distort Appearance uses it too.
PatternDef::offsets_covering overflowed on a region too large to count
in integers, as for a pattern scaled by 1e-20; it now gives no offsets.
Docs: docs/mcp.md, docs/control-protocol.md, docs/target-app-parity.md.
After the desktop save panel closes, the app adds the file type's
extension to a name typed without it (or with one the save does not
write). The panel asked only about the name as typed, so on Linux
typing `card` with `card.vectorcraft` in the folder wrote over
`card.vectorcraft` without a question. On Linux the save panel is the
XDG desktop portal, and its GTK and GNOME file choosers return the
name as typed. Windows and macOS add the extension themselves only
when the panel lists file types, and the panels of Document Info's
Save, the presets dialogs' Export…, the CSS Properties export, the
swatch and graphic style library saves and the Libraries panel's
export list none.
When the added extension names a file that exists, Save, Save As, Save
a Copy, Save as Template, the exports and the other writes through a
save panel now ask "“card.vectorcraft” already exists. Replace it?"
first (dialog `replaceFile`). OK runs the write again with the same
answer and replaces the file. Cancel writes nothing and shows the
dialog it was asked over again. The host reports whether a file exists
through a new `Services::exists`.
Docs: docs/control-protocol.md, docs/target-app-parity.md.
The toolbar listed the Touch Type tool (Shift+T), but choosing it did
nothing. The reference app's Touch Type tool moves, scales and rotates
one character of live type by dragging the handles of a box around it.
A click with the tool now selects one character and shows a box around
it. Dragging the box or its lower-left handle moves the character along
its line and up or down. The upper-left handle scales it vertically,
the lower-right handle scales it horizontally, the upper-right handle
scales it proportionally, and the upper-middle handle rotates it. A
character rotated by up to about 84 degrees follows the pointer too.
Near a quarter turn, a drag moves it only along its line. Escape
cancels a drag, and otherwise lets go of the character.
Each drag is one text.touchType command and one undo step. It sets the
character's baseline shift, scales and rotation, and moves it along its
line by changing its own kerning and the kerning of the character
before it, within the kerning range. The characters after it stay
where they are. Where the font kerned the character with its
neighbors, the two kerning values also keep the character in place
when it is only raised, scaled or rotated. The characters after it
stay where they were, or move by the change in its width when it is
scaled. The first character of a line and a right-to-left character
move only up and down. The type stays editable.
Docs: docs/mcp.md, docs/target-app-parity.md, docs/type-parity.md.
File > Place read text files only as plain `.txt`, and an RTF file
could not be placed. The reference app lists RTF among the formats it
places as text, keeping the file's font and style settings.
File > Place now reads RTF files, written from the Rich Text Format
Specification 1.9.1, and sets their text as area type the way a text
file is placed. Each run keeps its font from the font table, its size,
bold and italic, underline, strikethrough, all caps, small caps,
superscript or subscript (in Document Setup's proportions), baseline
shift and its color from the color table. Each paragraph keeps its
alignment, left, right and first-line indents, space before and after,
and direction. Table cells are separated by tabs and each table row is
a paragraph. Unicode characters, Windows-1252 and Mac Roman bytes and
the special-character control words are read, and of a \upr group's
ANSI and Unicode text, the Unicode text is read. Hidden text, pictures,
objects, headers, footers, notes, field instructions and other
destinations are left out, and unknown control words are skipped.
A file is read as RTF when it starts with {\rtf, whatever its name,
except a .txt file, which stays plain text. RTF files share the 4 MiB
limit of text files. Groups nested more than 512 deep share the
formatting of the deepest one read, the font and color tables keep at
most 4,096 entries, and a file with more than 100,000 paragraphs or
runs is refused. The Place dialog lists RTF files, and file.place,
file.place.info and file.place.queue report the format `rtf`.
The tests in tests_textimport.rs fail without the change, and
text_import_fuzz.rs places random RTF token streams.
Docs: docs/file-format-parity.md, docs/mcp.md,
docs/target-app-parity.md.
A Shift-click or Ctrl/Cmd-click on a brush in the Brushes panel used to
apply the brush, as a plain click does. It now selects brushes without
applying one. Shift-click adds the brushes from the last brush clicked
without Shift (else the highlighted brush) to the clicked one, in the
order the panel shows them. Ctrl/Cmd-click adds the clicked brush to the
highlighted brushes or removes it. The panel highlights the selected
brushes, and the bottom bar's Delete Brush and the panel menu's
Duplicate Brush and Delete Brush act on all of them. A plain click
works as before and ends the multi-selection, and so does selecting
other art, another document or another current brush.
brush.delete and brush.duplicate also take {names: [name...]}. Each
call is one undo step, and an unknown name fails before anything
changes. brush.duplicate answers {names} for a names list.
Docs: docs/mcp.md, docs/target-app-parity.md.
The reference app has a Blend panel, opened with Window > Blend, that
makes and edits blends in one place. VectorCraft had the Blend tool,
Object › Blend and the Blend Options dialog, but no Blend panel.
Window › Blend opens a panel with Make, the selected blend's Spacing and
Orientation, Reverse Front to Back, Reverse Spine, Replace Spine, Expand
and Release. Each control runs its Object › Blend command, so a change
of the selected blend's spacing or orientation shows on the canvas at
once and is one undo step. With two or more objects selected and no
blend, Spacing and Orientation set the options new blends start with,
as Blend Options does. Make is dimmed until two or more objects are
selected. Spacing and Orientation are dimmed until a blend or two or
more objects are selected, and the other buttons are dimmed until a
blend is selected. The Blend Options dialog's orientation toggles now
have names for screen readers.
Docs: docs/target-app-parity.md, docs/ui-parity.md.
The errors of the four blend spine commands (Move Spine Point, Reshape
Spine Segment, Add Spine Point and Delete Spine Point), which Direct
Selection, the Pen and the Anchor Point tool send, had no @msg row in
any catalog, so an error that reached the status bar, for example from
a Pen click, showed in English. The scan in
complete_languages_translate_every_message does not read messages
built with bad(), so that test did not report them.
Each of the 13 messages now has a row in the 16 catalogs that translate
every status and error message, in Bulgarian and in Simplified Chinese.
The rows that name the spine use each catalog's term for it. No English
string changes.
Docs: docs/target-app-parity.md.
A small closed loop whose fit collapsed to a single cubic (fit_closed),
or a piece returning to its own junction with fewer than two fitted
cubics (fit_piece), kept every positioned point as a straight segment.
A looser Paths collapses more loops, so it could give more segments
than a tighter one.
Such a loop now tries the fewest segments that still hold it, each only
if the result keeps an area (more than 0.25 px^2) and stays within
1.5 * max(tolerance, 0.5 px) of the points both ways: the fit at up to
three somewhat tighter tolerances (what a slightly higher Paths gives);
for a closed loop then two cubics split at its start and its farthest
point, then the outline simplified at the tolerance and finer. The
positioned points stay the last resort, as before.
Test (code-drawn): a_higher_paths_never_gives_fewer_segments sweeps
Paths 0/25/50/75/100 over 32 seeded 48 px images of soft blobs plus
pixel noise in Color and Black and White at Noise 1 and 3. Before, six
runs had fewer segments at a higher Paths (e.g. 125 -> 79 from Paths 0
to 25); after, none. One image (seed 16, Black and White: 140 -> 136
from Paths 0 to 25, with or without this change) has another cause and
is left out of the sweep.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018tBcdwNhptqtmX32JCGFzB
Where the pixel at a sharp corner falls to the other colour, the outline
cuts the corner in two one-pixel steps, and between two straight edges
of three cracks or more subpixel::outline makes those steps one point
(a notch). Whether the edges were long enough was measured on the whole
loop. An edge can run on past a junction in one region's loop while
the region across it, whose loop turns at that junction, sees a shorter
edge: one side made the notch and the other did not, so the shared
piece got different points on each side and the abutting curves left a
gap.
The notch now also requires the three cracks of each edge next to it to
lie in the piece (no junction among them), as the corner-pixel shift
correction above it already does.
Test (code-drawn): noisy_colour_images_leave_no_sample_uncovered traces
20 seeded 36 px images of soft overlapping blobs in four colours plus
pixel noise (Color, 5 colours, Noise 1 and 2, Paths 0/50/100) and needs
every one of five sample points per pixel to lie in some path: 93
uncovered before (seeds 9, 11, 13), 0 after.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018tBcdwNhptqtmX32JCGFzB
Layers::shift read a crack's sub-pixel shift from its inside pixel,
looking for a thin run on the inside before one on the outside. Where
both sides are thin (a one- or two-pixel sliver beside another, as in
soft gradients and noisy scans) the two regions of the crack got
different shifts, so their shared piece got different points and the
abutting curves left slivers that no path covers.
The shift is now the mean of the readings from both sides,
(s(c) - s(c flipped)) / 2, which is exactly antisymmetric in floating
point (Crack::flipped gives the crack seen from its other side).
Tests (code-drawn images):
- a_crack_between_thin_slivers_has_the_same_shift_from_either_side:
every crack's shift is the exact negative of its flipped crack's, on
a noisy ripple of 1-2 px columns at three thresholds.
- shared_boundaries::thin_slivers_on_both_sides_of_an_edge_leave_nothing_uncovered:
a 48 px noisy ripple in Black and White at thresholds 140/157/170 and
Paths 0/50/100, sampled on the shared-boundary check's 4x4 grid per
pixel: 482-1,228 uncovered points per trace before, 0 after.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018tBcdwNhptqtmX32JCGFzB
Flat Logo found its colours only from solid pixels (a uniform 3 x 3
neighbourhood). Line art on paper has no solid area but the paper's: a
stroke one or two pixels wide never fills one, so its colour was a handful
of pixels among tens of thousands of white and was cut as noise (under
0.1% of the solid pixels). The palette came back empty and trace_logo
returned an empty TraceResult, which imageTrace.make reported as a success
with paths 0, anchors 0, colors 0.
Cause: logo::palette::auto_palette had no way to find a colour that has no
solid area, and logo::pipeline::trace_logo treats an empty palette as
nothing to trace.
Fix: when the solid pixels give no colour but white, auto_palette reads
the colours from the strokes' cores instead (line_art_palette): the opaque
ink pixels at least as far from white as each of their four neighbours,
the middle of each stroke where it is least blended with the paper.
Two guards keep that fallback to line art:
- Ink is a colour at least 48 below white in one channel, and the page
must have ink on at least 1% of its opaque pixels. Light grain and a few
sparse marks on otherwise blank paper stay blank, as plain white does.
- The lighter shades along a stroke's anti-aliased edge are local maxima
of their own; a colour that lies on the line from white to a darker
colour of the palette (within 20) is a blend of that ink with the paper
and is folded into it (fold_fringes), the ink keeping its colour.
The rest of the pipeline (unmixing, the flat-colour check and its NotFlat
refusal) is unchanged, and images with a solid colour other than white
never reach the fallback.
Tests (code-drawn inputs):
- thin_line_art_on_paper_is_traced_not_blank: four 2.4 px dark strokes
and two red marks on white. Before: palette [], no paths. After: dark
and red paths along the middle of every stroke and mark.
- anti_aliased_coloured_lettering_has_only_its_inks: rings, slants and
bars 1.6 px wide in red, blue and green. Without the fold: six colours
besides white; after: exactly the three inks.
- sparse_marks_on_paper_are_not_promoted_to_colours: 24 grey 12 x 2 px
marks of 239, 223 or 191 on a 256 x 256 page (under 1% ink). Without
the guard: palette [white, #efefef]; after: empty palette, no paths.
- grainy_blank_paper_has_no_ink: paper grain 243-255 traces to nothing.
- edge_shades_fold_into_their_ink_and_other_colours_stay (unit).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018tBcdwNhptqtmX32JCGFzB
contour::trace_mask built a table of outgoing directions for every
pixel corner in a raster pass and then followed it. It now works out
each turn during the walk from the two pixels ahead of the current
corner: right if the one ahead on the right is outside, straight if the
one ahead on the left is, else left (so at a saddle it still turns
right and diagonal pixels stay apart). Loop starts are found by
scanning the pixels for a top edge on the outline, with one flag per
pixel for a top edge already walked. The walk is bounded by the grid
size, and a mask of the wrong length now gives no loops instead of
failing an assert_eq!.
The output is identical: the_walk_equals_the_stored_direction_walk
compares it with a test-only copy of the previous walk on all 65,536
4x4 masks and 400 seeded 23x17 masks.
Time (best of 7, release, 4 cores): the walk alone on 2000x2000 masks
19.1 vs 26.6 ms (smooth pattern), 81.3 vs 93.2 ms (12% random pixels),
223.1 vs 238.1 ms (50% random); whole 1000x1000 traces 73.1 vs 77.1 ms
(Black and White), 249.6 vs 251.7 ms (Color, 6 colours), 214.1 vs
218.6 ms (Color, 6 colours, Noise 1), same segment counts.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018tBcdwNhptqtmX32JCGFzB
A .ai or EPS opened from its editing data lost the opacity and blend
mode of its text objects. text_slot made the slot without the graphics
state's opacity and blend mode, which paths, meshes and images take, so
an Xy before a text object was lost. An Xy after it reached the slot,
but fill_slots then put type made from the text document in the slot's
place with only its visibility and name, and style_marker dropped the
transparency that a text object with an appearance carries on its look.
text_slot now takes the state's opacity and blend mode. fill_slots
gives made type the slot's opacity, blend mode, isolation and knockout.
The page's type gets them too, in place of its own transparency when
that is none, the slot's, or the slot's combined with that of the
objects around it; otherwise the page's type and the strokes around it
keep their own (a fill's, for example). When the type has transparency
and comes with other art (the strokes the page draws around it, or page
type that goes with it), a group of them takes the transparency. A text
object with an appearance takes the transparency from its look; where
the look draws more than the type and has transparency, the look and
the type go into a group that has it, and that group is not taken as
the look of a later object. The import fuzz's .ai tokens now include a
commented text object, 1 /FreeUndo , and a group's state marker
(6 () XW).
Docs: docs/mcp.md, docs/target-app-parity.md.
Shift pressed while dragging a segment had no effect on the curve:
Direct Selection's curved-segment and spine-segment drags, the Anchor
Point tool and the Pen with Alt sent the pointer offset alone, and the
segment bent freely. The reference app keeps the handles in the
perpendicular direction and of equal length while Shift is held.
path.reshapeSegment and object.blend.spine.reshapeSegment take
perpendicular (default false, which keeps today's behavior). With it,
both of the segment's handles stand at right angles to the line between
its anchors, on the same side and at the same length, and the point at
t lies as far from that line as the moved point does. A smooth anchor
at either end stays smooth, as it does in a drag without perpendicular,
and a segment whose anchors coincide bends as it does without
perpendicular. With Direct Selection, Shift pressed once a curved
segment's drag has started sends perpendicular; Shift held at the press
still toggles the path's selection, as before. Direct Selection's
spine-segment drag and the Anchor Point tool's segment drag send Shift
as perpendicular too, so the Pen with Alt does as well. A straight
segment dragged with Direct Selection already moves at 45° steps with
Shift.
Docs: docs/mcp.md, docs/target-app-parity.md.
Shift+Enter in a panel or Control bar field ended its editing, as Enter
does. The reference app's default keyboard shortcuts list "Apply value
and keep text box active" under Use panels and "Apply a value and keep
focus in edit field" under Transform panel, both Shift+Enter on Windows
and Shift+Return on macOS.
- A focused single-line field of the shared widgets (num_field,
mixed_field, plain_field, range_field, spin_field, spin_field_auto,
spin_plain, text_field, exact_text_field, opt_field and hex_field)
takes Shift+Enter before its text box reads it, applies the text as
Enter does and keeps the focus. In the next frame it shows the value
as it now is, all selected, so typing replaces it.
- A text box of several lines keeps Shift+Enter as a line break.
- While a dialog is open, Shift+Enter applies the field's value and
confirms the dialog, as before. The app sets widgets::set_dialog_open
at the start of each frame.
Docs: docs/target-app-parity.md.
Smart Blur's Edge Only and Overlay Edge modes marked a pixel when one
of its four nearest pixels differed from it by more than Threshold,
and Radius changed only the blur. The reference app's documentation
says that Radius sets the size of the area searched for dissimilar
pixels.
A pixel is now on an edge when one of the pixels the search reads
within Radius differs from it by more than Threshold and its
channels, alpha included, add up to at least that pixel's. The search
picks its neighbors as the Normal blur does, over the whole pixels of
Radius and within it, and always reads the four nearest pixels.
A color step gives a line about as wide as Radius on the side with
the larger sum, and a Radius of one pixel or less gives the same
edges as before. The edge modes' reach() includes the blurred copy
that a wide Radius reads.
On a 2,000 by 2,000 px image, Edge Only takes 0.49 s at Radius 10 and
0.34 s at Radius 100, and Normal takes 0.73 s and 0.39 s. Normal
draws the same pixels as before.
Docs: docs/target-app-parity.md.
The PostScript interpreter that reads EPS and PostScript files had no
CCITTFaxDecode filter. Reading a file whose image data went through it
stopped at an error naming the filter, and the file opened from a
fallback, such as its preview or the art drawn before the image, or
did not open.
CCITTFaxDecode now decodes CCITT Group 4 (ITU-T T.6) and Group 3
one-dimensional and mixed (ITU-T T.4) data to rows of 1-bit samples.
It reads K, Columns, Rows, EndOfBlock, EndOfLine, EncodedByteAlign and
BlackIs1 (PLRM 3rd ed., §3.13; ISO 32000-2, §7.4.6). With EndOfBlock
the data ends at its EOFB or RTC regardless of Rows, and the RTC of
mixed Group 3 data, six EOLs each with a tag bit, is found by the
filter. Damaged data keeps the rows decoded before it, and damage in
the first row is an ioerror. Columns outside 1 to 65,535 and Rows
below 0 are a rangecheck, and data that would decode to more than
256 MiB is refused. hayro-ccitt 0.3.0, already in Cargo.lock through
hayro-syntax, becomes a dependency of vectorcraft-eps.
The tests encode their images in the test, with the fax crate's
Group 4 encoder and T.4 run codes (a dev-dependency, already in
Cargo.lock through tiff), and fail without the change. The import fuzz
tests read a Group 4 and a Group 3 image coded by hand from the T.4
and T.6 tables, read the same file with random damage, and add the
filter and its parameters to the hostile programs.
Docs: docs/file-format-parity.md, docs/target-app-parity.md.
The EPS reader read PostScript by any name, but the Open and Place
dialogs hid .ps and .epsf files, and a .ps or .epsf file whose header
is only %! could not be opened or placed ("not a format VectorCraft
reads"). The reference app lists .eps, .epsf and .ps under Open and
Place.
- A read-only PostScript format (ps) joins the format table. A .ps
file opens and places through the EPS reader and reports the
ps format. document.export does not write .ps; file.print does.
- .epsf is an EPS extension.
- OPEN_EXTS and PLACE_EXTS list epsf and ps, so the dialogs, the CLI
help and the MCP open_file description list them.
UNASSOCIATED_EXTS lists them too: the macOS bundle declares neither.
- A .ps or .epsf file whose header is only %! now pastes as EPS when
copied in a file manager, as one with a %!PS header already did, and
clipboard.importEps reads PostScript whose header is only %!, which
it rejected before.
Docs: docs/mcp.md, docs/file-format-parity.md (.ps row),
docs/target-app-parity.md.
A double-click on a ruler guide with the Selection, Direct Selection or
Group Selection tool did nothing, whatever keys were held. A guide could
be released only with View › Guides › Release Guides.
With the Selection, Direct Selection or Group Selection tool, a
double-click with Cmd/Ctrl and Shift held now runs view.guides.release
on the ruler guide under the pointer (index) or, when no ruler guide
is there, on the guide path under it (ids). Nothing is released while
guides are hidden or locked, and a double-click without both keys does
what it did before.
Docs: docs/mcp.md, docs/target-app-parity.md.
The Layers panel menu had no saved-selection items, and the Select
menu was the only menu that listed saved selections. The reference
app's help also names a Save Selection menu in the Layers panel that
lists the saved selections and holds Save Selection, Update Selection
and Edit Selection.
The Layers panel menu now has a Save Selection submenu after Locate
Object. It holds Save Selection…, Update Selection and Edit Selection…,
which are enabled and run as in the Select menu, and below them the
active document's saved selections. Choosing a name runs select.recall
with that name, as the Select menu's list does.
Docs: docs/mcp.md, docs/target-app-parity.md.
Native files remember the grid like the saved view, so reopening restores it.
Toggling it is still view state: no undo step, document not marked modified.
Fixes#1968
With several artboards selected, the Artboards panel menu's Duplicate
Artboards copied only the active one, and Copy and Cut with the
Artboard tool took only the tool's active artboard.
artboard.duplicate, artboard.copy and artboard.cut now take `indices`,
and without `index` or `indices` they act on the artboards the
Artboard tool has selected. Duplicate Artboards passes the selected
artboards and selects the copies. Duplicate and Cut are one undo step
each, and Paste adds every copied artboard back, with its art, in one
undo step. The copies keep their places relative to each other,
together right of the last artboard. A cut fails when it names every
artboard.
Docs: docs/mcp.md, docs/target-app-parity.md.
Labels that the UI looks up through a variable or a context (#1259) had
rows only in the Bulgarian, Czech, Greek, Spanish, Finnish, French,
Polish, Brazilian Portuguese and Swedish catalogs. In German, Hungarian,
Italian, Japanese, Korean, Russian, Ukrainian, Vietnamese and both
Chinese, the UI showed most of them in English: the file dialogs' filter
names, the status bar's render and UI timings, New Document's preset
name, and the contributor tooltips with their plurals. The Printer mark
type list showed English "Roman" in each of those languages except
Hungarian, and in Japanese, Korean, Vietnamese and both Chinese it
showed the language's name for Japanese.
Those catalogs get the rows, each next to its related entry. German also
gets rows for the 3D gizmo's light, which the UI showed as "Hell"
(bright), and for Grain's Regular type, which the UI showed in English.
complete_catalogs_translate_render_site_labels requires the rows those
catalogs get and the light's row in every complete catalog, and checks
German's Regular row.
Docs: docs/localization-parity.md, docs/target-app-parity.md.
file.saveAsTemplate kept a name whose extension named another format
and wrote the template into it: {path: "card.png"} wrote .vctemplate
JSON to card.png, and card.pdf typed in the desktop save panel got it
too. file.saveAs {format: "ait"} answered that document.export writes
other formats, but document.export does not write .ait either.
file.saveAsTemplate now replaces another format's extension (card.png
is saved as card.vctemplate, or card.ait for format: "ait") and adds
the extension to a name whose extension names no format. The desktop
panel adds the template's extension to any name that does not end in
a template extension (card.pdf.vctemplate), as Save As adds its own.
Save and Save As still refuse ait, and the error now names
file.saveAsTemplate.
Docs: docs/mcp.md, docs/target-app-parity.md.
effect.apply and effect.setParams stored any string as the color of
Drop Shadow, Inner Glow, Outer Glow and Shift to Color. When the
effect could not read it, it drew its default color with no message,
so a typo such as #ff00 in the Drop Shadow dialog turned the shadow
black.
The commands now refuse such a color before they edit, with
"`color` must be a #rrggbb color", and the document and its undo
history stay as they were. Hex colors, the five named colors,
[r, g, b] and color objects still pass. The check runs in the
commands, so the app, command.batch steps and replayed journals get
it too. The effect dialog's OK shows the error and keeps the dialog
open, with Preview on or off. Saved documents open as before.
Docs: docs/mcp.md (effect params).
A style captured from text left out the attributes that the text's
JSON skips at their defaults, such as the Hyphenation options, the
Composer, tab stops, Auto kerning and the character alignment.
Redefining a paragraph or character style from text set
back to the defaults therefore did not reach the text using the
style: a paragraph that had the old Hyphen Limit or Single-line
Composer from the style kept it. setAttrs dropped a default value it
was given in the same way.
New and Redefine now capture every attribute of the selected text,
and charStyle.setAttrs and paraStyle.setAttrs keep a default value.
When a style changes, text whose value equals the old definition's
takes the new value, a default included. Local overrides on text that
uses a style other than the Normal styles stay, and an attribute that
setAttrs leaves out still leaves the text's value alone. The Character
Styles and Paragraph Styles panels compare the attributes a style holds
with the same attributes of the selection. A paragraph style made from
type with Auto alignment no longer shows "+" on that type.
Two existing behaviors change. Applying a style made with New or
Redefine from text, without Clear Overrides, now also sets the
attributes that text had at their defaults, such as Auto kerning, no
OpenType features, the Every-line Composer, no tab stops and the
default Hyphenation options. Redefining [Normal Paragraph Style] or
[Normal Character Style] from text gives every paragraph or run that
uses it those attributes too, so local values of them on that text are
reset.
Docs: docs/target-app-parity.md.
Batch over data sets names each output file after the document and the
data set, as the reference app's File + Data Set Name option does. It
joined that name onto the destination folder as it was, and a data set
name can hold any character. A data set named a/b failed to save, and
one named a/../../x wrote its file outside the destination folder. On
Windows, one named c\d also failed to save, and one named e:f left an
empty file.
The data set's name now goes through file_part, which Export for
Screens uses for its file name prefix and a format row's suffix, so
output file names have - in place of /, \ and : from a data set's name.
When two data sets then give the same name, compared without case, the
later ones get -2, -3 and so on, as artboard file names do, through a
distinct_names helper split out of unique_file_names. This covers the
folder destination and Override Save/Export.
Docs: docs/mcp.md, docs/target-app-parity.md.
The Character panel's Small Caps, Superscript and Subscript buttons
were disabled, with "(on the roadmap)" tips, and always drawn off, and
the panel menu's items of the same names were disabled, although
text.setFormat and text.setRangeStyle already set smallCaps and
position. The reference app applies all three from the Character
panel menu.
The three buttons and menu items now turn small caps and the
superscript or subscript position on and off for the selected type, or
for the text the Type tool has selected, in Document Setup's
proportions. Turning Superscript on turns Subscript off, and turning
Subscript on turns Superscript off. They show as on when the first
selected character has the attribute, as the All Caps, Underline and
Strikethrough toggles do.
docs/mcp.md did not mention smallCaps and position; it now describes
them for text.setFormat and text.setRangeStyle.
Docs: docs/mcp.md, docs/target-app-parity.md.
An object whose art style had a Drop Shadow with Darkness (a csrc
other than 0 in the shadow filter's dictionary) came in as its drawn
look, with a warning that its appearance isn't editable yet. With a
Color shadow instead, the same object comes in as itself with its
fill, stroke and shadow.
The reader now reads such a shadow as Drop Shadow with shade
"darkness" and darkness set to the filter's dark value, a percentage
kept within 0 to 100, or 100 when the value is missing or not a
finite number. The shadow's other keys are read as for a shadow with Color,
and the object comes in as itself wherever it would with a shadow with
Color. A shadow with Color comes in with the same params as before.
The import fuzz's .ai tokens now include csrc and dark.
Docs: docs/target-app-parity.md.
Each style run was shaped on its own, so restyling one character, even
only its fill, dropped the font's kerning pairs at the new run
boundaries. A fill change on the V of 40 pt "AVAV" moved it 0.56 pt and
moved the characters after it 1.12 pt. Manual kerning on a character
also dropped the font's kerning with the character before it.
Adjacent runs that differ only in fill, stroke, underline,
strikethrough, overprint, character style name, leading, baseline
shift, rotation, Character Alignment, kerning, or tracking (as long as
both values leave standard ligatures on or both turn them off) are now
shaped together, and each glyph takes those attributes from its own
run. The font's kerning pairs apply after characters with automatic
kerning: a manual value replaces the pair after its character, and the
pair before it stays, except before a character that All Caps or Small
Caps sets as two letters (ß as SS). A change of font, size, scale or
OpenType features still shapes the characters apart, and a ligature or
a mark still does not join characters in different styles.
Docs: docs/target-app-parity.md.
SVG export wrote each style run of a line of live text as one <tspan>
and positioned the text again only after tabs and, on justified lines,
after word spaces. A viewer then spaced the characters inside the
<tspan> with the font's own advances, so it lost the letter gaps of a
Justify All line without word spaces, the gaps between the characters
of a justified Japanese line and the spacing that a Mojikumi Set
changes. On justified lines, the textLength of horizontally scaled type
also took in the justification space after it, so a viewer stretched
the glyphs into that space.
On a line that is justified or set with a Mojikumi Set, text_lines now
lays the text out again as point type without the Mojikumi Set to get
the advance a viewer gives each character. Each character that the
layout draws away from where a viewer would draw it starts a new
positioned <tspan>, and textLength and line anchors use the viewer's
widths. The last character of a horizontally scaled <tspan> takes its
width without the font's kerning against the next character. Other
lines, and lines written with Fewer <tspan> Elements, are written as
before.
Docs: docs/target-app-parity.md.
The Simplified Chinese catalog had 20 `@msg` rows, against 714 to 718
in each catalog listed in COMPLETE_MESSAGES, so with the interface in
Simplified Chinese most status-bar and error messages showed in
English.
- zh-hans.tsv gains the 694 missing `@msg` rows in one block with a
header comment on its conventions: no final 。 unless the English
ends with a period, a space between Chinese and adjacent Latin
text, numbers or placeholders, and 、 between list items. The rows
use the catalog's existing terms (画板, 图稿, 编组, 描边, 色板, 实时上色,
图像描摹, 增效工具 ...).
- zh-hans joins COMPLETE_MESSAGES, so
complete_languages_translate_every_message and the narrower message
tests fail when a zh-hans message row is missing, and a new message
needs a zh-hans row.
- The registry comment, docs/development.md (which also gains Greek,
already in COMPLETE_MESSAGES), the zh-hans row of
docs/localization-parity.md, and the interface-languages line of
docs/target-app-parity.md (which also gains Czech, Greek, Hungarian,
Swedish and Finnish) list Simplified Chinese among the languages
that translate messages.
docs/type-parity.md and the Set type row of docs/roadmap.md still
listed the Hyphenation… options as missing, although the Hyphenation
dialog (#2310) sets them for each paragraph.
The type-parity row is now partial and names Words Longer Than, After
First, Before Last, Hyphen Limit, Hyphenation Zone and Hyphenate
Capitalized Words. Justification… stays missing. The Set type row of
roadmap.md lists only the justification options as missing. Its score
is unchanged. docs/target-app-parity.md and docs/gaps.md are left as
they are.
TIFFs that store each channel as its own plane (PlanarConfiguration 2)
did not open, place or relink when they were CMYK or grayscale with
extra channels: the image decoder refuses planar CMYK, and the readers
for CMYK inks and for extra samples returned nothing for planar files.
The two readers now read every plane through the tiff crate and put
each pixel's samples together. A planar CMYK TIFF keeps its ink
amounts and is stored as a CMYK TIFF with each pixel's inks together,
at the file's resolution. Grayscale TIFFs with an alpha channel or
other extra channels, and CMYK TIFFs with alpha, read as their chunky
versions do.
A TIFF whose header claims more strips or tiles than the tiff decoder
can count is refused as too large to draw. The decoder multiplies
those counts without a check, which panics in builds with overflow
checks.
Part of #1466.
Docs: docs/file-format-parity.md, docs/target-app-parity.md.
A syllable drawn with several glyphs that share one cluster, such as a
pre-base i matra and its consonant (कि), had a caret stop at the edge
of each glyph. A click just right of the matra, in the left half of
the syllable, put the caret after the syllable. byte_in_line now takes
one extent for each cluster, and type on a path compares the click
with the middle of the cluster, from the start of its first glyph to
the end of its last.
New tests with an installed OpenType Devanagari font check that क्ष
forms one conjunct glyph, that र्क draws the reph after its base, that
कि draws the matra before its consonant, that Ka with a nukta matches
the precomposed क़, and that each syllable is one cluster with the
arrow keys stopping at its ends. They print SKIPPED when no such font
is installed.
Docs: docs/localization-parity.md (Hindi row), docs/target-app-parity.md.
dataset.import read only commas between cells and refused a file whose
first header cell wasn't `name`. A value that didn't match its
variable's kind, and a repeated data set name, were reported with the
row's record number, which is not the file line after a blank line or a
quoted line break. A data set name over 255 characters gave no line.
- When the header row, read with tabs between cells, has a tab and no
comma outside quotes, tabs separate the cells. A quoted cell may then
hold tabs and commas.
- Without a `name` column every header cell is a variable, and the rows
take the first free names Data Set 1, Data Set 2 and so on, as Capture
Data Set and blank name cells do. A first header cell `name`, in any
case, still holds the data set names.
- Value errors, a repeated data set name and a name over 255 characters
give the file line the row starts on.
- The Variables panel's Import Data Sets… dialog also lists .tsv files.
- import_fuzz.rs imports random text, random bytes and mutated exports,
comma- and tab-separated, into a document with bound text, visibility
and graph variables, selects up to three of the imported data sets in
turn, and exports every data set.
Docs: docs/mcp.md, docs/target-app-parity.md.
file.print, print.preview and print.setup took print settings only as
settings, so an agent had to read a preset with print.presets.list and
copy its settings into the call. They now take preset, the name of
[Default] or of a saved print preset in any case. The preset's settings
are used instead of the document's, settings applies over them, and
print.setup saves them with the document as one undo step. An unknown
name is a parameter error that lists the presets. The app's file.print
with params passes preset on to the engine.
Docs: docs/mcp.md, docs/target-app-parity.md.
With a gradient stop selected, an Alt-click on a color swatch painted
the inactive Fill or Stroke, as it does with no stop selected. The
reference app's Default keyboard shortcuts page lists "Apply swatch
color to active (or selected) color stop: Alt-click swatch in the
Swatches panel" for the Gradient panel.
Now, when the active Fill or Stroke is a linear or radial gradient with
a selected stop, an Alt-click on a color swatch or a tint swatch
recolors that stop in one undo step, and the stop stays selected. The
stop changes as a drop of the same swatch on the Gradient panel's ramp
changes it (drop_color and dropped_link in panels/gradient.rs): it
becomes fully opaque, and a global or spot color or a tint stays
linked, so editing the swatch recolors the stop.
With a freeform gradient, the selected point takes the color through
paint.freeform.setPoint. With no stop or point selected, Alt-click
paints the inactive Fill or Stroke as before. The Fill and Stroke chip
popovers show the same swatches and behave the same.
Docs: docs/control-protocol.md, docs/target-app-parity.md.
The controls that the shared widgets draw had no name in the
accessibility tree: icon, flat, primary, secondary and split buttons,
chips, toggles, checkboxes, radio buttons, tabs, sliders, the angle
dial, the Reference Point's squares, the spinner and preset arrows and
the searchable combo boxes. Screen readers announced each clickable one
as an unnamed element of unknown role, and a disabled icon button was
not in the tree. The Color Guide's library button, the Color Themes
Save button and the status bar's artboard buttons also had no name.
Each helper in widgets.rs except color_field now gives its response a
role, a name and its state. The name is the text, label or translated
tool tip the helper shows. The spinner, the preset arrow, the angle
dial, the harmony wheel and the Reference Point's squares, which show
none, are named "Step up or down", "Presets", "Angle", "Color wheel"
and Top Left to Bottom Right. A toggle, checkbox, radio button, tab or
Reference Point square reports whether it is on, an icon button reports
that it is pressed while it is lit, the angle dial reports its angle,
and a searchable combo box reports the text it shows. A checkbox or
radio button drawn without text is named by the label beside it. The
Tools panel's tool buttons, flyout rows and floating strips are named
by their tools' tool tips, with the active tool pressed. The title
bar's column toggle, the tear-off bars, a floating strip's close box
and the Color, Gradient and None buttons are named by their tool tips.
The status bar's First, Previous, Next and Last Artboard buttons now
have tool tips, and the Fill and Stroke chips' tool tips are
translated. The six new strings (the four artboard tips, "Color wheel"
and "Step up or down") have rows in every catalog.
split_button, chip_button, color_slider, channel_slider and the dialog
form's slider_field and slider_rail keep their arguments. New _named
variants beside them take the arrow's tool tip, the chip's tool tip or
the slider's name, and every caller in the tree uses them.
Docs: docs/target-app-parity.md.
PDF presets are kept one file each in a PDF Presets folder, ICC profiles as copies in a Color Profiles folder, next to the preferences; the desktop app and vectorcraft-cli use them. Adds color.removeProfile.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The MCP server read one line, ran it and answered it before reading the next,
so notifications/cancelled was never seen during a call and every request
(even ping) waited behind a running export.
- vectorcraft_doc::cancel: a thread-local cancel token and progress sink that
long work polls. Checkpoints: PDF pages, PDF flattening, geometry bake,
raster-effect flattening, drawing objects, raster artboards, and before any
file is written. A cancelled command returns EngineError::Cancelled and
changes nothing.
- Server::serve reads on its own thread: notifications/cancelled drops a
queued request or stops the running one (no reply, as the MCP spec says),
engine-free requests (ping, the list methods, prompts/get) are answered at
once, a request that would wait behind a call running for 2 s is refused
with -32003 naming that call, and _meta.progressToken gets
notifications/progress (page n of m). Queue and output are bounded.
- The remote backend stops waiting for the app when the call is cancelled.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
dotCooCoo