When the browser dropped the page's WebGPU device or WebGL context, the
editor kept running but nothing it drew reached the screen: Chrome and
WebKit showed a frozen, stale canvas and Firefox a blank page, while
edits and Save went on.
- ui-egui: new `graphics` module. `GraphicsLoss` carries a loss report
from any thread (wgpu's device-lost callback, the canvas's
`webglcontextlost` event) to the UI thread. `VectorcraftApp::
graphics_lost` writes Data Recovery copies of the modified documents at
once. The app now notices when the host hands it a new egui context and
uploads everything again (fonts and theme, the canvas, place cursor
thumbnails); every per-thread texture cache (layer, symbol, swatch,
style, asset and glyph thumbnails, previews, font samples, Navigator,
Flattener Preview) is a `TexCache` that empties for a new context.
Without this, the moved app drew stale or wrong textures and egui
panicked on its missing font families.
- web: the shell watches both loss signals; on a loss it puts a fresh
canvas in place of the dead one and starts the same `WebRunner` again,
handing over the same app (documents, undo, panels), with services made
for the new context and a status-bar note. If no new graphics can be had
(or after three losses) it removes the stale canvas, lets go of the
recovery copies and says to reload.
- engine: `recovery::leave` lets go of a session's area but keeps its
copies, so the reloaded page offers them at once instead of after the
heartbeat goes stale.
- desktop: eframe can't give a window a new device, so a lost device
(driver reset) is logged, the copies are written and the status bar says
to save and restart; closing then skips the Save Changes question it
can't show.
- Tests: loss reporting across threads, `TexCache` emptying, the app
uploading the canvas, a cached preview, theme and fonts again in a new
context (fails without the fix), copies written on loss, and
`recovery::leave` offering copies at once with both lock and heartbeat
stores. docs/development.md (web section) and ROADMAP updated.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A live rectangle stored four corner radii but drew them all with the largest,
and the corner widgets always rounded all four corners, so one corner could not
be rounded on its own.
- geom: rectangle_with_corners draws each corner with its own radius and kind
(Round, Inverted Round, Chamfer); a sharp corner is one anchor, a cut one two.
Uniform corners keep the exact rounded-rectangle anchors.
- doc: LiveShape::Rectangle gains `kinds` (saved only when a corner isn't round;
older files read as round), plus helpers mapping anchors to corners.
- engine: object.setLiveShape takes `corners` and `kind`; without `corners` it
edits the Direct-Selected corners, else all four, and keeps the selected
corners selected as their anchors change.
- tools: with Direct Selection, only the selected corners show widgets and a
drag rounds just those; Alt-click cycles the corner kind; double-click opens
the new Corners dialog (also `ui.corners`).
- ui: Properties/Transform radius fields show and set the selected corners;
CSS export keeps per-corner radii and skips cut-in corners.
- docs: mcp.md (Live Corners), control-protocol.md (Corners dialog), ROADMAP.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
An object Illustrator writes as a fill of its outline left open and then a
stroke of it closed imported as two objects: a filled path and a separate
stroke-only path. A fill and a stroke of the same outline (equal once every
subpath is closed) now become one path with the fill and a live stroke
(width, cap, join, miter limit, dash, paint), taking the stroke's outline.
Layers: a group whose view state is off (/Usage /View /ViewState /OFF, with or
without an /AS view event) comes in as a hidden layer with its art, as groups
off in the default configuration already did. The configuration's /Order
nesting (inline or as its own object, labels skipped) makes sublayers, which
take their top-level layer's colour; nesting depth, entries and groups are
capped. Splitting a page's art by group is no longer quadratic, so thousands
of layers import quickly.
The 6D art pen (rotation-aware brushes) is a separate feature, now on the
ROADMAP gap list.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Shape Builder only used the filled areas of the selected paths, so a
stroke-only line or open path across a shape did nothing. Like the
reference app's Shape Builder, open paths now act as cutting edges:
- pathops::shape_builder: with an open path in the selection, the
regions are the Live Paint planar map's faces (every path an edge), so a
line across a shape splits it and lines can enclose a region of their
own; it also returns each open path cut where other paths meet it. Closed
shapes alone, more than SHAPE_BUILDER_MAX_SEGMENTS segments, or a map
that can't be built fall back to the filled-area arrangement.
- Filled open paths count as closed (an invisible closing edge) where that
encloses an area; unfilled open paths and straight lines stay edges.
- shapeBuilder.merge: merging or deleting regions on both sides of a line
piece removes that piece (the line is split, its other pieces kept); an
area only lines enclose takes the current fill and the front-most
object's look. Alt (erase) within `tolerance` of a line piece deletes it.
Returns `lines` (pieces deleted).
- shapeBuilder.regions also lists the line pieces.
- The tool shares BuilderMap/Touched with the command, highlights the line
piece under the pointer while Alt is held and passes its tolerance.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
`command.batch` rolls a failed batch back through each document's interaction, which restored the
art, the selection, the current layer and isolation but not the highlighted Layers panel rows. A
batch whose `layer.setCurrent` step ran before a later step failed left that layer's row
highlighted, so a following Collect in New Layer collected the layer instead of the selected art.
The failed batch isn't journaled, so replaying the journal gave a different document: the
`journal_replay_reproduces_document` property test failed whenever it drew that sequence.
The interaction now keeps the highlighted rows too, which also puts them back when a drag is
cancelled.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A heartbeat alone can't tell a crashed tab from one whose timers stopped in
the background, so another tab could discard (or restore) the copies of a
still-live document once its heartbeat went stale.
- Recovery stores without locks can now announce an area for as long as the
app runs (RecoveryStore::announce/announced). The web shell implements it
with the browser's Web Locks (navigator.locks through js_sys::Reflect): a
tab holds a lock named after its area, which the browser releases only when
the tab is gone. Lock snapshots are refreshed every 10 s (once before the
first frame); a snapshot over a minute old counts every area as held.
- An area is gone only when no announcement holds it and its heartbeat is
stale. Where the store can't tell (no Web Locks), the owner repairs: each
heartbeat (also its first frame after a pause) and each recovery save write
again the copies missing from its area, and a heartbeat newer than the one
another tab wrote when taking the area over marks it running again.
- MemoryStore::announcing models Web Locks in tests; engine and UI tests cover
the issue's sequence. Native folder locks are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- The em box top that Top-to-Top Leading asks of every glyph's line metrics parsed the font's
cmap and vertical metrics per glyph on every layout: each face now reads it once.
- The Paragraph panel menu shows Top-to-Top / Bottom-to-Bottom Leading with the East Asian
options (Preferences › Type), as the Mojikumi Set does, and Reset Panel goes back to
Bottom-to-Bottom Leading, the default.
- Saved as `leading_model`, like the other paragraph attributes in files.
- Simplified Chinese (landed since) translates the two menu items.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Japanese layout measures leading from the top of one line's ideographic
em box to the next (the right side, in vertical type), not from baseline
to baseline. The Paragraph panel menu now offers both:
- `ParaStyle.leading_model`: RomanBaseline (as before, the default) or
EmBoxTop. With EmBoxTop a line's leading is the space below it — the
next line's em box top is this line's plus its leading, and the next
baseline hangs from there by that line's tallest em box — and area
type's first line touches the top of the frame (its em box top on the
frame's top, at least Area Type Options' minimum). With baseline
leading a line's leading stays the space above it.
- Lines of one size are exactly as far apart either way; the models
differ for mixed sizes and for the first line of area type (fonts
whose ascent isn't the em box top: Hiragino's nearly is).
- The em box top comes from the font's vertical metrics (#343), else
0.88 em. Written only when EmBoxTop.
- `text.setFormat {leadingModel: "romanBaseline"|"emBoxTop"}`; menu items
Top-to-Top Leading / Bottom-to-Bottom Leading (checked as set); catalog
rows in ja and zh-hant.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Rotate, Scale, Reflect and Shear tools answer Enter with their
dialog, but the shortcut handler only applied what a tool asks of the
UI for the keys it claims, so the request was dropped and Enter did
nothing. It now applies it for Enter, and consumes the key when a
dialog opened, since a dialog takes Enter as its OK in the same frame.
The Selection, Direct Selection and Group Selection tools open the
Move dialog the same way, on Enter and on a double-click on their
toolbar button (`tool.options`), as Object > Transform > Move does.
With nothing selected neither opens anything.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Opening a PDF whose fonts aren't installed keeps the text live in the fallback font (as
Illustrator substitutes and highlights missing fonts), and opening said so, but every export
then wrote the fallback font silently ("warnings": []), so a headless convert re-set the type
without a word.
- Exports that draw type now say so in their warnings, naming the missing families and the
fallback font: PDF (outlined or real text, also .ai and Save As), EPS, EMF/WMF, raster images
and SVG with outlined or embedded fonts. Live SVG text, DXF, text and native files name the
font itself and don't warn. Replacing the font (Find Font) clears it.
- The fonts a document uses (text.fonts' scope, Package's font list) come from one helper.
- textAs: outlines already keeps the glyphs the PDF draws (its embedded font program, installed
or not): now covered by a test with an embedded serif font that isn't installed, checked
against the installed font's and the fallback's outlines.
- docs/mcp.md and ROADMAP describe the warning.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A PDF's DeviceCMYK image opened as RGB (the interpreter's pixels as a PNG),
so every export wrote it as DeviceRGB or ICC RGB, silently. Images now keep
their CMYK samples, as CMYK colours already did:
- Model: CMYK images are blobs holding CMYK samples, a four-component JPEG
(Adobe's inverted samples) or a Deflate CMYK TIFF. vectorcraft-doc's cmyk
module reads their ink amounts (ImageBlob::is_cmyk, ::cmyk, capped at 64
megapixels) and writes the TIFF (ImageBlob::cmyk_tiff).
- PDF import: DeviceCMYK, CalCMYK and four-component ICC images keep their
samples. A JPEG with Decode [1 0 1 0 1 0 1 0] is kept byte for byte; any
other is decoded (8 or 16 bits, its Decode array applied) and stored as a
CMYK TIFF. Masked CMYK images and 1/2/4-bit ones open in RGB with a warning.
Placing a CMYK TIFF keeps it CMYK too.
- PDF export: CMYK images are written in CMYK (DeviceCMYK, or ICC-based with
the CMYK profile when colours are tagged: PDF/X-3, PDF/X-4, PDF/A). A CMYK
JPEG neither resampled nor recompressed passes through unchanged; others are
resampled in CMYK and compressed as ZIP or as a CMYK JPEG (jpeg-encoder,
which also replaces the ZIP fallback and its warning for images converted to
CMYK). Output conversion treats them like CMYK colours: Preserve Numbers and
PDF/X-1a keep their numbers, Convert to Destination converts them to another
CMYK profile or to RGB through a 17^4 lookup table (cms::CmykLut, now also
behind the renderer's ink table). A CMYK image that can't be decoded is
reported in the warnings.
- Canvas: CMYK images show through the working CMYK profile, as CMYK fills
do, and paint their own inks in ink planes (CMYK raster exports, CMYK
blending). SVG export writes TIFF images as PNG.
Illustrator's Output › No Conversion keeps colour data as it is
(https://helpx.adobe.com/illustrator/using/pdf-options.html).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A PDF-compatible .ai paints the editor's page (an opaque white rectangle the
size of the page) before its first optional-content layer and outside every
layer. It came in as art: an extra "Page 1" layer with a page-sized white path
behind the artwork, which hid white art and gave SVG/PNG/PDF exports an
opaque background.
For pages with the editor's private data (the way the importer already tells
a .ai from a PDF), the first art outside every layer, painted before any
layer, is dropped when it is a plain fill-only rectangle in an opaque,
non-spot white, with normal blending, equal to the crop box (0.5 pt
tolerance). Plain PDFs keep such a rectangle as art (as the reference app
does when opening a PDF), and so does anything else: grey, smaller or larger
rectangles, strokes, translucency, white inks, fills over the layers.
The white-cover test the soft-mask reader used is now a shared helper.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Adds a Spanish catalog (crates/ui-egui/src/i18n/es.tsv) covering the whole
interface like Japanese and Traditional Chinese: every menu string, every
tl! literal, the hint-bar fragments, plurals and the runtime strings the
other complete catalogs carry. It is written from the meaning of the
English labels in neutral, international Spanish (Latin American usage)
with the vocabulary users of vector illustration software know (Mesa de
trabajo, Muestras, Trazo, Relleno, Trazado, Pluma, Selección directa,
Buscatrazos, Motivo, Preajuste); no other application's localisation was
copied. Like Japanese it keeps the product, OpenType, the built-in
workspace names and the perspective grid preset names in English.
The language is registered with complete_menus (so the completeness
tests enforce it) and the one/other plural rule; every es-* system locale
(es-ES, es-MX, es-AR, es-419 ...) resolves to it.
Tests: es locale tags, Spanish vocabulary and plurals, Spanish letters in
the bundled UI fonts, and a new check that catalogs written with spaces
keep the leading/trailing spaces of joined fragments. ROADMAP, README and
docs/development.md list the language.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Drop the workspace [patch.crates-io] and restore windows-link's crates.io
source and checksum in Cargo.lock, so every normal build keeps the
checksum-verified crate. windows7.ps1 now applies vendor/windows-link with
--config, checks that only windows-link's lock entry moved, builds and tests
with --locked, checks the patched crate is the one in use, and restores
Cargo.lock afterwards.
- windows7.yml runs on demand and weekly instead of on every PR, pins every
action by commit SHA, and leaves the dependency check and host tests to the
script (-Test) so the Cargo arguments live in one place.
- ROADMAP, README, NOTICE and docs: experimental, unsupported, not verified by
maintainers.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With Alt down at the press, the four transform tools opened their
dialog at once, so an Alt-drag did nothing: no transform and no copy.
They now decide on release: a click made with Alt still sets the
reference point and opens the dialog, and a drag made with Alt
transforms a copy about the reference point, as it does in
Illustrator.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The Up/Down stepping of #357 now serves every numeric field through one
helper: num_field (lengths, in their unit) and plain/mixed fields
(degrees, percent, counts, at their precision: a count ignores the
Ctrl/Cmd tenth), as in Illustrator's panels and dialogs. After a step
the new value is selected, so typing replaces it. The select-all code
the dialog focus and focus/double-click already duplicated becomes one
helper, and the tests move to tests_numfields.rs with its field harness
(which can now run with the window unfocused). ROADMAP.md notes it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With Alt held, a press on a path's stroke selects the whole path and
an Alt-drag copies it, as before #363 and as Group Selection does in
Illustrator; only a plain press on a segment selects its anchors and
reshapes it. Records the segment drag in ROADMAP.md.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Tab is the Panels shortcut (show/hide all panels), but egui's own
focus navigation read the same Tab press at the start of the frame and
moved the keyboard focus on to a field. The shortcut handler returns
early while a field has the keyboard, so after one Tab a second Tab,
tool letters and Ctrl chords did nothing until a click or Esc, and in
the Type tool the letter typed after a Tab was lost. When no field has
the keyboard, the handler now cancels the focus move for that frame;
in dialogs, the command palette and panel fields, Tab still moves
between fields.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Panel number fields (num_field: X, Y, W, H, corner radius, stroke
weight, font size...) only parsed their text on Enter or focus loss,
so the arrow keys did nothing. With the field focused, Up/Down now
step the value by one of the field's unit (Shift: ten, Ctrl/Cmd: a
tenth), from the value shown, and apply it at once, as in the
reference app.
The field's focus is read from egui's memory, as its highlight and its
typing do: Response::has_focus also needs the window to report
keyboard focus, which the running app's window doesn't always do.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Any hit on a path's stroke took the "select the whole path" branch, so
dragging a segment moved the whole path. Now a press on the stroke of
a path that isn't selected as a whole selects the segment's two
anchors (unless they already are), and the drag bends the segment
with path.reshapeSegment when it's curved, else moves the selected
anchors (a straight edge moves, the rest stays), as in the reference
app. The fill, Shift-click and a path selected as a whole keep their
behaviour.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Finder, Open With, a drop on the Dock icon and `open -a` failed with "VectorCraft cannot open
files in the PDF document format":
- The release Info.plist declared no document types, so Launch Services refused the files before
the app saw them. Info.plist.in now declares every format File › Open reads (the engine's
OPEN_EXTS; a test keeps them in step): VectorCraft owns its own documents and templates and is an
alternate for the others, so it takes over no other app's files. `cargo xtask bundle` now fills
in the same template instead of its own, shorter plist.
- macOS delivers such files as open-document events, not arguments. winit 0.30's application
delegate has no `application:openURLs:`, so AppKit fell back to NSDocumentController, which
refuses every file of a non-NSDocument app (the error in both issues). The new
`open_documents` module adds that method to the delegate's class when the app is about to finish
launching (before AppKit delivers launch files), queues the paths and wakes the UI, which opens
them with `io::open_path` like command-line files; errors go to the status bar too.
New direct macOS dependencies `objc2-foundation` 0.3 and `block2` 0.6 were already in Cargo.lock
(through winit, muda and rfd). The module allows `unsafe_code` for the Objective-C runtime call.
Fixes#354
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The web reads dropped files asynchronously, and the place inbox only kept
the drop point, so a file dropped on document A was placed into whatever
document was active when its bytes arrived. A drop on a canvas now carries
place::DropAt (the document's uid, the point, embed), and drop_files makes
that document active again before placing; when it has closed meanwhile,
nothing is placed and the status bar says so. The native drop path is
synchronous and goes through the same code, so it is unchanged in effect.
Test: a drop pending while another document opens lands in the first one
at the drop point; one whose document closes is not placed anywhere.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Simplified Chinese is a complete catalog, so the strings added the same day need rows: Delete
Anchor Points, Edit Selection and the 25 Saved Selection slots, Paragraph Direction and the
Mojikumi Set. Remove Anchor Points reads 移除锚点, apart from Delete Anchor Points (删除锚点), as in
Traditional Chinese.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
BFlat
imachus
theironchef
njmueller