The PR added `checked_anchor_from_json` next to the lenient
`anchor_from_json` and dropped path.rs's import of the latter, which
`path.appendAnchor` still called, so the engine didn't compile.
`anchor_from_json` is now the strict parser itself (handles through
`finite_numbers`), used by path.create, path.setAnchors and
path.appendAnchor alike: the Pen's appended anchor with a malformed
handle fails too instead of losing the handle. Tests: a non-boolean
`closed` and a malformed appended handle change nothing. rustfmt.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A rejected Stroke panel edit with nothing selected leaves the next
object's stroke as it was; an accepted one sets its weight and dashes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Select › Reselect repeated the last selection command, but a Select Same
command needs a selected object to compare with, so after Deselect it
failed with "nothing selected". Reselect now remembers the objects that
were selected when the command ran and, with nothing selected, takes the
reference from those still in the document.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A run (vectorcraft-cli run, the MCP command_batch tool) had no way to use
an id an earlier step returned. A parameter string that is exactly
$N.path or $last.path now takes that value from step N's result
(0-based), e.g. {"id":"$1.id"}; $$ starts a literal $, and a
reference to nothing fails the step. Objects given to a command are now
used as given: an ids element or id that isn't an object id (or an
unresolved reference) fails instead of the command acting on the
selection. text.setRangeStyle styles the selected type object when no
id is given, as text.setStyle does.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With the Selection tool, a click (no drag, no modifier) on one object of a
selection of several makes it the key object (select.key); a click on the
key again lets it go. The key object is drawn with a thicker outline,
Align aligns to it (the Align panel shows Align to Key Object while there
is a key, and choosing Selection or Artboard lets it go), and
object.align without `to` aligns to the key when there is one, so the
Control bar's and the Properties panel's Align buttons do too. With a
spacing, Distribute Spacing keeps the key object still and spaces the
others from it, and it reports too few objects instead of indexing an
empty list.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- The inspect_document tool (and the control channel's document.inspect)
take {depth?, childLimit?}, so MCP clients get the skeleton without
going through run_command; the tool description points at the
drill-down path.
- document.find rejects a filter of the wrong type instead of dropping
it (which widened the search); one opt_u64 helper parses limit, depth
and childLimit.
- docs/mcp.md spells out find's matching rules; control-protocol.md
and ROADMAP updated; MCP and engine tests cover the new paths.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
document.inspect takes the same {depth?, childLimit?} slicing as
document.node (a depth-0 read is the layer skeleton), and document.find
{name?|kind?|text?, limit?} locates nodes across the tree with their
ancestor paths and a total. docs/mcp.md documents the drill-down recipe;
the CLI stdio sweep covers both.
`depth` / `childLimit` without `summary: true` used to return the whole
full-JSON subtree silently; it is now an invalid-params error. docs/mcp.md
describes the slice parameters and childCount, the engine test covers
fractional values, unsliced reads and limits past the tree, and the CLI
stdio sweep reads a sliced group through run_command.
Co-Authored-By: Claude Opus 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>
A new sublayer was named "Layer {node_count}", counting every object, so
an empty document's first sublayer was "Layer 1" like its parent and a
document with five objects got "Layer 6". Top-level layers counted only
top-level names and could repeat a sublayer's. Document::next_layer_name
now counts every layer, sublayers included, and skips any layer's name;
layer.newSublayer uses it too.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Between paths, the Pen now adds an anchor where a click lands on a
segment of a selected path and deletes the anchor a click lands on,
with the add and delete cursors while hovering, as in the reference
app. The ends of a selected open path still continue it. Shift held,
or General > Disable Auto Add/Delete (a stored preference nothing read
until now, passed to tools as ToolContext::auto_add_delete), starts a
new path instead.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A type node's fill, stroke and strokeWidth now come from its first run, as
paint.proxies and its strokeOptions already did; fills and strokes of the
type object itself are reported apart as objectFill, objectStroke and
objectStrokeWidth.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Shipped code no longer unwraps, expects or panics: every one of the 124
unwrap()/expect()/panic!/unreachable! sites outside tests now handles the
failure (Result/Option, ?, let-else, defaults) and reports errors to the user.
- Lints: workspace clippy denies unwrap_used, expect_used, panic, todo,
unimplemented and unreachable; clippy.toml allows them in tests, and
integration tests, examples and testkit opt out explicitly.
- Safety net (vectorcraft_engine::guard): a panic in a top-level command rolls
the document back and returns EngineError::Internal; a panicking tool is
reset; the MCP server answers with an internal error and keeps serving; the
UI frame and each control request are guarded.
- APIs that hid failures now return them: Rendered::to_png/to_jpeg/to_webp
return Result (JPEG/WebP errors used to write empty files),
FontDb::face returns Option, Region::top returns Option.
- Fuzzing: engine/tests/import_fuzz.rs feeds garbage, hostile and mutated SVG
and PDF through import → render → export (30k cases locally, no panics).
- Docs: AGENTS.md/CLAUDE.md rule, docs/development.md "Robustness" section,
README pointer and a PR template checklist.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
New preferences start with Scale Strokes & Effects off, as in Illustrator:
scaling an object keeps its stroke weights, dashes and effect distances unless
the option (or `strokes: true`) is on. Preference files that already store the
option keep their value.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The engine unit tests share one process, and two colour-management tests
switched the process-wide Color Settings (the working RGB space to the wide
gamut profile, Black Point Compensation off) for a moment. Any test that
showed or rendered a colour at that moment saw it through the wrong profile:
clicking_an_image_samples_its_pixel read #008181 instead of #008080.
- Move the two tests that change the Color Settings to their own test
binary (crates/engine/tests/color_settings.rs), so no unit test can see
non-default settings.
- The image-sampling test compares the stored colour, not its display hex.
- every_command_survives_empty_params toggles the process-wide proof view
(Proof Colors, Overprint Preview, Separations), so it now takes the
colour-management test lock like the tests that read that view.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
