Review follow-up (CodeRabbit): the two older G2 bullets counting 46 and 49
filters were superseded by Conté Crayon (57/57) and are folded into one
current bullet; the G2 row in the ranked table and the entry's estimate now
both say 4–6 h, as ROADMAP.md and roadmap.md do. ROADMAP.md's revision
history gains its 2026-10-22 row.
- One budget for the whole read (layers, pixels, characters, table
cells), shared with embedded documents and charged also when their
reading fails; the first reading is dropped before the second.
- Tables: at most 65 536 cells; lines and cell formats that don't
match the grid are left out instead of shifted.
- Paragraph formats matched to runs by binary search; font names
longer than 256 bytes are dropped.
- Overlays: one work budget per import; encoded originals decoded
within their layer's size; table cell outlines recoloured too.
- Engine: master page copies capped, each page's master found in
linear time, line gaps looked up once per font, the working CMYK
table built once, objects past MAX_COORD left out.
- Embedded content stays with its node; pixel masks under artistic
text are read whole and include stroke reach; gradient overlays on
hidden layers are kept.
- Render: contexts and raster effect images stay within
MAX_RASTER_SIDE per side (vello overflowed near u16::MAX; found by
the new fuzz target).
- Tests: affinity tests/import_limits.rs, feature samples in
import_fuzz.rs, shared builders in engine/tests/support/.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
- Strokes set to None no longer import as black 1px strokes.
- Underline and strikethrough are kept.
- Frame text starts at cap height; lines hanging out of a frame make it taller;
mixed-alignment artistic text lines up within its longest line; line breaks
inside a paragraph get no paragraph spacing.
- Table cell text taller than its cell is centred and stays visible.
- Outline effects on stroked shapes are drawn around the stroke.
- CMYK pixel layers go through the document's embedded CMYK profile; without
one, opaque layers stay CMYK.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The Print dialog's Media list only had Letter/Legal/Tabloid and the ISO A
and B papers, so Japanese papers had to be entered as a custom size, and
"B4"/"B5" were ISO 216 (250 × 353 and 176 × 250 mm) although Japanese
printing uses the JIS sizes (257 × 364 and 182 × 257 mm, JIS X 6310).
Media gains JIS B4, B5 and B6 next to the ISO ones, and はがき
(japanese postcard, 100 × 148 mm), 往復はがき (japanese double postcard,
148 × 200 mm), 封筒 長形3号/長形4号 (japanese envelope #3/#4, 120 × 235
and 90 × 225 mm), 角形2号/角形3号 (kaku #2/#3, 240 × 332 and 216 × 277 mm),
四六判 (yonroku-ban, 127 × 188 mm) and A6, all in millimetres at 72/25.4.
b4 and b5 keep their ISO 216 dimensions so print settings and presets
saved before the Japanese papers existed keep their paper; the dialog
labels say "B4 (ISO)"/"B5 (ISO)" beside "B4 (JIS)"/"B5 (JIS)".
A paper is named by its JSON id or by its dialog label, any case and
spacing ("b5 (jis)", "B5 (JIS)" and "b5jis" all mean the JIS B5), so
agents and the control channel need not know the id spelling; an unknown
name is still refused. Media is written out instead of through the
choice! macro to carry that lookup.
The print.setup/file.print parameter docs and docs/mcp.md list the new
papers; the Print dialog picks them up from Choice::ALL with no UI code
change, and the Japanese labels are translated in ja.tsv (はがき,
往復はがき, 長形3号, 長形4号, 角形2号, 角形3号, 四六判).
Tests: japanese_paper_sizes checks every added size in points, that the
ISO B4/B5 values did not change, name resolution and id/label round
trips, and that each paper prints a PDF page with its own MediaBox size.
Validation: cargo fmt --all -- --check clean; cargo clippy
-p vectorcraft-pdf --all-targets -- -D warnings clean; cargo test
-p vectorcraft-pdf -> 330 passed, 0 failed (328 lib + 2 integration).
Squashed from #2533 onto #2525 (Smudge Stick, Sponge, Watercolor): both sets in PIXEL_EFFECTS (45), the catalogs and the docs, whose counts now read 45/57.
ROADMAP: a progress row and status line for the community PRs of 2026-10-19; the reproducible-bytes and inch-ruler rows appeared twice. gaps.md: G6 held four interleaved versions of its Missing line; it now says what is left (a scripting surface). target-app-parity.md: three copies of the Type evidence bullet, merged into one with every sentence. type-parity.md and ui-parity.md: a row and a shorter copy of it.
Labelling to a hundred and twenty-eighth left the left ruler unreadable: it
spells a label one character per line, 8.5 pt apart, and at zoom 50 the step is
1/64 with labels 56.25 pt apart while "0.015625" needs 61.5 pt. At zoom 100 it
is "0.0078125" in the same space.
The label ladder stops at a sixteenth, as the report says Illustrator's does,
so zooming in past that leaves the labels where they are and only the ticks
between them grow denser -- down to a hundred and twenty-eighth, which is the
32nds, 64ths and 128ths the report expects to appear unlabelled. Every label is
at most the six characters of "0.0625", which fits the 50 pt the ladder leaves.
Zooming in on an inch ruler stepped through 0.2, 0.1 and 0.05 of an inch:
`Unit::ruler_step` picked from one decimal ladder for every unit, and the
rulers divided whatever it returned by ten. A tenth of an inch is not a mark an
inch ruler carries (#2827).
Below an inch the steps are binary fractions now, 1/2 down to 1/128, and
`Unit::ruler_minor` gives the unlabelled ticks between two labels: an eighth of
the step on an inch ruler, so every tick is a binary fraction of an inch and
they stay about as dense as the tenths elsewhere. An inch and wider keeps the
shared ladder, and the metric units, points, pixels and picas keep their
tenths. Ruler-tick snapping follows the ticks that are drawn, which moves an
inch document's Shift-drag of a guide onto the same grid.
`ruler_label` took its decimals from `(-step.log10()).ceil()`, which writes
1/16 as "0.06" and 1/8 as "0.13" -- wrong values, not just rounded ones. It
uses the fewest decimals that write the step exactly, and drops trailing zeros
as `Unit::number` does, so a sixteenth-inch ruler labels the inch `1` rather
than spelling `1.0000` down six lines of the vertical ruler.
Two review points on #2880.
The reproducibility claim needed qualifying. PDF encryption draws its AES
initialisation vectors and the file `/ID` from a random seed, which
`SOURCE_DATE_EPOCH` deliberately does not reach, so a PDF exported with a
password never repeats however the variable is set. The docs and the ROADMAP
line say so now.
The byte-identity test passes an explicit `created`, so it never exercised the
clock-or-variable fallback. `pinned_or` takes the environment and the clock as
arguments, and the new test shows the variable is preferred, that a blank,
negative or non-numeric value leaves it to the clock, and that a host with
neither has no date to record. The environment is injected rather than set: a
test that changes this process's environment is `unsafe` under edition 2024 and
would reach every other test running beside it, which is why `cursor_theme`
takes a lookup closure too.
PNG, PDF, EPS, legacy `.ai`, packaged zips and saved `.vectorcraft` files
record the wall-clock time they were made in, so the same commands run twice
gave different bytes although the pixels matched (#2880):
PDF /CreationDate and /ModDate D:20261018155816 vs D:20261018155817
PNG the `Creation Time` tEXt entry
.vectorcraft one byte of the stored date
SVG carries no time, and already repeated
`metadata::output_time` is the time a file records: `SOURCE_DATE_EPOCH` when it
holds a non-negative whole number of seconds, else the clock, as the
reproducible-builds specification defines it. A date a command is given
(`file.new {created}`, a save's `modified`) still wins over the variable, and a
blank, negative or non-numeric value is ignored so that an export is never
refused over the environment.
`metadata::now_unix` keeps the real clock and now has exactly two callers, both
of which must not be pinnable from the environment: the recovery heartbeat, and
the seed PDF encryption draws on.
While an export runs in the background (Background Export), a dialog shows
what is being exported and how far it has got: artboard by artboard for a
raster export of several artboards, busy otherwise. Cancel (also Esc) stops
it and writes no file; Continue in Background leaves it to the status bar.
It opens after 0.4 s and never over another dialog.
Built on the shared vectorcraft_doc::cancel (#2546): fileio::progress::Progress
wraps its CancelToken and a progress sink, so a cancelled export stops
drawing and keeps no half-drawn pattern tile or effect raster.
- ui: background jobs carry their Progress; export_progress draws the
dialog; UI commands file.cancelExport and file.exportInBackground.
- i18n: the strings in every language with complete menus, Danish, Dutch
and Indonesian included; "{n} more exports waiting" is a plural (trn).
- docs: ui-parity, gaps (G12) and ROADMAP rows.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dt49z1ubK9x1bPq51APgz3
Add opaque RGB light color and up to ten independently adjustable lights to
Revolve, Extrude & Bevel, Inflate and Rotate. Keep light selection, Add/Delete,
selected-light gizmos and preview transactions on the existing command path.
Preserve legacy white-light Inflate polygons, paint composition and seam joins,
including the 0.05 seam minimum. Use shared smooth normals for colored Inflate,
and batch solid-fill RGB channel paths behind one surface clip. Keep source
paths editable and SVG/PDF output vector artwork.
Rebase onto main f7de3d6a, retaining both sides of progress docs and all 22
catalogs. Shade freeform points once through the shared paint mapper. Document
release measurements and a reproducible CLI fixture, including the remaining
source-gradient cost and the one-light timing tradeoff.
Validation: full cargo xtask ci passed all 10 steps; 7,978 tests passed, zero
failed and 11 ignored. Four legacy white-light PNGs match unchanged main byte
for byte. Three paired release bench/perf repetitions completed without busy
warnings; all six standard perf runs passed every budget. Two-light solid
Inflate improved from 136.2 to 96.8 ms/frame at 1200x1200, default eight threads.
Three original Artistic filters on the pixel pipeline, each with a
catalogue row, the generic live dialog, translations and tests.
- Cutout: colours softened by Edge Simplicity (less with Edge Fidelity),
split into Number of Levels steps per channel and painted as flat
pieces in their average colour; slivers along softened edges take the
nearest large piece's colour.
- Poster Edges: channels posterized to Posterization + 2 levels, black
lines on the edges of any channel and of the outline, widened by Edge
Thickness and darkened by Edge Intensity.
- Neon Glow: dimmed greys screened with a glow colour from the light
areas (positive Glow Size) or the dark ones (negative), as strong as
Glow Brightness.
All keep coverage and work in document units at any Document Raster
Effects resolution. Photoshop filters 33/57; cargo xtask ci passes all
9 steps.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Node editing as the public manual describes it: Direct Selection's anchor squares (hollow
or filled), Reshape Segment over a curved segment, Group Selection's plus (also Direct
Selection with Alt), the arrow hollow while a drag snaps, and Alt-copy arrowheads. The
Anchor Point caret, the Curvature pen and Free Transform's side-shear arrows (ShearH,
ShearV) are drawn in code. Lasso, the freehand, symbolism and other tools return cursors
of their own that borrow their old look until drawn (Cursor::stand_in).
crates/tools/src/cursors.toml lists every Cursor value, when it shows and its hotspot,
for whoever draws them; `cargo xtask cursors` (in ci) warns about a missing, unknown,
repeated or incomplete entry and never fails. The busy cursor is listed as missing (G1).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The Effect menu had no Rasterize. Object › Rasterize replaces the
object with an image, so its paths can no longer be edited.
Effect › Rasterize… (`rasterize`) is a live effect with Object ›
Rasterize's options: color model, resolution (the document's raster
effects resolution by default), white or transparent background,
anti-aliasing, Create Clipping Mask and Add Around Object. The canvas
draws the object as square pixels over its bounds, and its paths, type
and other art stay as they are. The effects above it in the stack,
shadows included, are part of its image, and shadows below it are cast
by the image. When Rasterize is an object's only raster effect, Expand
Appearance and PDF and SVG export give the image Object › Rasterize
gives with the same options.
Object › Rasterize now measures the art as drawn, with its geometry
effects, brush art and raster effects, so a Transform effect that moves
the art no longer leaves the image empty. Its Create Clipping Mask
clips the image to the art as drawn. The canvas no longer panics
drawing a pixel effect such as Gaussian Blur on an object about 65,500
screen pixels wide: the raster of such an object is at most 32,768
pixels a side, and the object is drawn into it at a lower resolution.
Docs: docs/mcp.md, docs/effects-parity.md, docs/target-app-parity.md,
ROADMAP.
Adds Swedish (Svenska) as an interface language: 4770 strings, written from the English source.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Keep surviving drawable runs separate when Delete or Backspace removes an open-path anchor. Preserve appearance, selection and one-step Undo.
Co-authored-by: Codex <noreply@openai.com>
When entering text edit mode via double-click from a selection tool
(Selection or Direct Selection), pressing Escape now returns to the specific
tool used before double-clicking (rather than always falling back to
Selection).
Additionally, when editing text while in isolation mode (such as when drilling
into nested groups with the Selection tool), pressing Escape exits text editing
while keeping isolation mode active; a second Escape press is required to exit
isolation mode.
- Record the previous tool in TypeTool on double-click from canvas.
- Return to prev_tool on ToolKey::Escape, consuming the key in shortcuts when claimed.
- Add unit test in vectorcraft-tools verifying Escape returns to prev_tool.
- Add integration tests in vectorcraft-ui-egui verifying Escape returns to Direct
Selection and that isolation mode persists until a second Escape.
- Update docs/target-app-parity.md and ROADMAP.md.
With the Direct Selection tool ("white arrow", shortcut 'a'), double-clicking
a text element (either standalone or nested inside a group) now switches to
the Type tool and immediately positions the caret at the clicked character,
matching the behavior previously introduced for the Selection tool in #284.
- In DirectSelectionTool::pointer, detect double-click hits on NodeKind::Text
when !self.group and emit Action::SwitchTool("type").
- Group Selection (group: true) retains its existing group cycling behavior.
- Add unit test in vectorcraft-tools verifying tool switch on both standalone
and grouped text.
- Add headless egui integration test in vectorcraft-ui-egui verifying tool switch,
caret positioning, and inline editing without triggering shortcuts.
- Update docs/target-app-parity.md and ROADMAP.md.
Taha-Mahmoodi