Flat Logo counted the anchors of every fitted outline and hole before it
left out the ones with no area, so a trace whose kept paths fit the
anchor limit could still be refused as too complex. The limit now counts
the anchors of the subpaths that go into the result.
The limit on outlines per color also counted holes, which are cut only
with Ignore White on. Holes now count toward it only when they are cut.
`tests_alternate`'s Arabic tests look up installed fonts by name
(`FontDb::face`), which loads each one into the global database and so
changes its generation, without `tests_fontchange::adding_fonts()`.
When that lands in the first frame of
`tests_charstroke::character_stroke_outlines_are_kept_between_frames`,
the second frame lays the type out again and expands the stroke again,
and the test fails (3 runs in 30 on macOS, where Arial and Geeza Pro
are installed; where no Arabic font is, those tests return early and
nothing changes). They now hold the lock, and its documentation says
looking up an installed font counts as changing the fonts.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Found while testing crates/svg/src/import/files.rs:82 (percent_decoded)
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Found while testing crates/engine/src/steps.rs:74 (parse)
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
With a Live Paint group in the selection, the Shape Builder took every
visible path in the group as an input. Unfilled faces became regions of
their own and came back from a merge as invisible paths. An edge that
closes on itself, such as the outline of a single rectangle, became a
second shape over its face: a click on the face gave the merged shape
the edge's look, a stroke with no fill.
The Shape Builder now takes the group's visible filled faces as shapes
and its visible stroked edges as open lines; a closed edge is opened at
its first anchor. A merge replaces the group with the merged shape, the
filled faces it did not use and the edges, except the edges between
merged faces, as it does for other selected paths and lines. An area
that the edges enclose around an unfilled face is still a region. The
Live Paint commands still collect every face and edge.
Holding Alt over a selection the Shape Builder had already mapped built
the planar map a second time, to add the outline edges that erasing
hits. A merge, and a Live Paint Bucket fill of a plain selection, also
dropped the map, so a gesture that left the document unchanged was
followed by another full build.
The Shape Builder now builds its map with the outline edges the first
time, and both tools keep the map until the document, its revision or
the selection changes.
When the inputs have more than 128 segments, a hover with the Shape
Builder or the Live Paint Bucket now builds the map on another thread
and highlights no region until it arrives, instead of holding up the
frame; the canvas checks for it each frame. A press waits for the map,
and the headless MCP host waits for it after a move. The web build has
no threads and builds the map at once. The Live Paint Selection tool's
hover no longer builds a map of the selection, which it did not use.
Object › Flatten Transparency replaced a frame of threaded text by its
outlines with Convert All Text to Outlines, and by flat regions or an
image where the frame was transparent. The thread's re-flow then added
the frame's text to the story, as it does for a frame that was deleted,
so the other frames held that text a second time while the outlines,
regions or image drew it.
A frame that Flatten Transparency outlines or flattens now leaves its
thread with the text it shows, and the other frames keep the text they
hold, as with Create Outlines and Expand. When the thread's last frame
is flattened, the text it has no room for goes after the text of the
last frame that stays. A thread left with one frame is no longer a
thread, and undo restores the thread.
Docs: docs/mcp.md, docs/target-app-parity.md.
A merge removed every piece of an open line that had a merged region on
both sides. The end of a line that stops inside a shape has the same
region on both sides, so a click or a drag on the shape, or deleting it
in erase mode, cut the line back to where it enters the shape.
A piece is now removed only when it runs between two different merged
or deleted regions. A line that crosses a shape still splits it, and
merging the two halves still removes the piece between them.
A drag tested points spaced by a step that grew with the selection (at
least 0.25 pt), and the tool's highlight tested points 2 screen pixels
apart. A region narrower than the step, such as a thin overlap, fell
between samples: it kept its piece of each source, the merge left strips
behind, or the drag failed with "no region under the pointer" while the
highlight showed the region. Zoomed out, the highlight could miss a
region the merge then took.
The drag now finds where each segment crosses a region outline and tests
the middle of every stretch between crossings, so every region it
crosses is picked. The tool's highlight and shapeBuilder.merge use the
same picking, and Shaper's scribble picks the faces it crosses the same
way. In erase mode the line pieces and edges are still found within the
tolerance, and a segment that crosses one also tests the crossing point.
Shaper's scribble found the strokes to erase at points at least 0.5 pt
apart, so with a tolerance below 1 pt it could pass over a stroke
without erasing it. It now also tests where the scribble crosses each
stroke and points tolerance / 2 apart.
`open -a` and PowerShell's `&` return before the app opens its control
port, and `vectorcraft-cli mcp` tries the port only once, so the docs
now say to wait for the VectorCraft window before starting the MCP
client. The `--connect` failure is documented as the full line the CLI
prints, `vectorcraft-cli: cannot connect to 127.0.0.1:7979: ` followed
by the system's connection error.
A paragraph split between two threaded frames took its first-line
indent again at the top of the next frame, and a list paragraph drew
its bullet or number there again. Each frame is laid out on its own,
and its first paragraph kept every paragraph attribute there.
A text object now marks a frame whose first paragraph continues from
the frame before (first_para_continues, not saved). The thread re-flow
sets the mark where each frame's share starts inside a paragraph and
measures each share with it, and a restored document reads the marks
from its saved text. Area type and type on a path set the first line of
such a paragraph as a wrapped line: no first-line indent and no bullet
or number. The paragraph after it takes the next number. Release
Selection, Remove Threading, Convert To Point Type and copies of a
frame clear the mark, so the first paragraph there has its indent and
bullet or number again. When Create Outlines, Expand or Expand
Appearance takes frames out of a thread, a frame that stays keeps the
mark only while it is threaded and the frames before it end inside a
paragraph. A frame whose mark changes lays out its first line again,
and when it is still threaded, its thread re-flows. When the outlined
frame is the last one, the text it has no room for is measured as the
frame lays it out: without the first-line indent when it continues a
paragraph, and with its numbers counted on from the frames before it.
The frame before it now takes exactly the text the last frame has no
room for. Duplicates, copies dragged in the Layers panel, the copies
that Release to Layers (Build), Release Opacity Mask and Edit Opacity
Mask make, and Draw Inside's clipping copy then have their cached
bounds computed again, as paste already does for pasted type.
Docs: docs/mcp.md, docs/target-app-parity.md.
pdf_real_text_draws_svg_glyphs_as_art_and_keeps_their_characters counted
every line containing Tj or TJ. The file ID in the trailer and the XMP is
base64 derived from the timestamp, so about one run in a hundred it spells
one of them and the test saw two extra lines (main CI, 5c1185dd). Only
lines that begin a text object count now.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Shape Builder, Live Paint Bucket and Live Paint Selection keep the
selection's region map between pointer events, keyed on the document's
address and the selected ids. Every edit replaces the document, and after
an undo followed by an edit a later document can be allocated at the
address of the one the map was built from. The key then matched a map of
an earlier revision, so a drag across the art merged nothing and a Live
Paint Bucket click filled nothing.
The key now also holds the tool context's revision, which moves with every
edit, undo and selection change.
A Paintbrush sample dropped for being under a pixel from the last point
left the ring computed from the pointer, while the release keeps the last
point; the ring now follows that point (CodeRabbit).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A Pencil or Paintbrush stroke released within Close Paths When Ends Are
Within of its start closes, but the cursor gave no sign of it. While the
pointer is where a release would close the stroke, the cursor is now
PencilClose (PaintbrushClose for the Paintbrush), the crosshair with a
ring, as Draw straight lines with the Pencil tool describes:
https://helpx.adobe.com/illustrator/desktop/draw-shapes-and-paths/draw-shapes/draw-straight-lines-with-the-pencil-tool.html
The check is the one the release makes, on the end the stroke would
take (snapped for the Pencil). A straight line (Shift or Alt held at the
press), a stroke that continues a selected path, a stroke that redraws
one and a smoothing drag show no ring.
Docs: docs/target-app-parity.md, docs/ui-parity.md.
Without a loader, check() scanned the script once, so a function using a
variable the global code assigns after it (`function g(){ f.write('x'); }
var f = new File('a');`) missed file.write, while the loader path found it.
The no-loader path now runs the same fix-point, keeping a scan that finds
nothing new as the report scan, so a script that includes nothing gets the
same report with and without a loader.
Fixes#4011
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Found while testing crates/doc/src/xml_id.rs:32 (from_object_name), decoded by escaped() at :72
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Within about a minute of its window no longer being visible, macOS App
Nap lowers the app's threads to the lowest priority and coalesces its
timers. On a loaded machine, control requests then took seconds to
answer, and the 60 s timeout reply arrived up to 26 s late.
When the control channel is on (--control or VECTORCRAFT_CONTROL_PORT),
the app now holds an NSProcessInfo activity (user initiated, allowing
idle system sleep, latency critical) for as long as it runs, so App Nap
does not apply. The activity does not keep the Mac awake.
docs/control-protocol.md states this in the paragraph on the 60 s
timeout.
Move the declaration pass, the significant-token helpers and the feature
recognisers into analyze/declared.rs, sig.rs and features.rs without
changing their code, and replace the duplicated line-break and directive
dispatch of scan and declarations with LineBreaks::read.
Part of #4014
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
FeatureId and OutsideId replace the string ids the analyzer passed around.
One table each gives the id, the need and the report order, so
catalog::FEATURES and catalog::need() are gone. Feature.id and the engine
JSON are unchanged; OUTSIDE is generated from the same table.
Fixes#4013
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Work out the unescaped text and the declared names of each file once and
share them between the scans, keep included files as bytes and decode
them on demand, check the cheap tokens before unquoting in the '=' case,
and stop allocating a String per recorded chain.
Fixes#4010Fixes#4012
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
dotCooCoo