The raster-effect cache key now includes the setting, so a renderer that
draws the same art with it on and then off no longer reuses a drop
shadow or other raster effect made with the other setting.
The Navigator panel menu's Panel Options… is enabled and opens a dialog with
the view box color (a preset or a custom color; Light Red, today's color, is
the default) and Draw Dashed Lines as Solid Lines, off by default, which draws
dashed strokes without their dashes in the panel's thumbnail only. The
options are kept with the UI state, and ui.navigatorPanelOptions sets them.
RenderOptions::solid_dashes draws dashed strokes as solid lines, including
type strokes and the silhouettes of stroke effects.
layer.setProps, layer.new, layer.newSublayer and livePaint.gapOptions read
an 8-digit hex color such as #ff000080 as #ff0000 and set it without its
alpha. They now refuse it with the error a bad color already gets, and
change nothing. #rgba was already refused. #rrggbb, #rgb, a preset index
and a preset name are read as before.
`Color::from_hex` reads `#rrggbbaa` and keeps only its first six digits, so
a `color` set to `#ff000080` through `ui.dialog.set` was applied as
`#ff0000` without an error by Tile Edge Color, Live Paint Selection Tool
Options, Live Paint Bucket Options and Gap Options.
- A new `checked` next to `resolve` in `dialogs/tile_edge_color.rs` gives
the existing "is not a colour" error for an 8-digit hex color, and
otherwise reads the color as `resolve` does. `resolve` is unchanged, so
the swatches the dialogs draw with it show such a color by its RGB as
before.
- Each of the four dialogs' OK checks the color with `checked` first, and
leaves the dialog open with nothing changed. Tile Edge Color and Gap
Options already refused `#rgba`, which `from_hex` does not read. Live
Paint Selection and Live Paint Bucket used to ignore a color they could
not read, such as `#f008`, and apply their other fields; their OK now
refuses it.
- With Preview on, Gap Options sends no color for a `#rrggbbaa` value, so
the groups keep their own color, and the previewed gaps draw in black.
Docs: the params text, docs/mcp.md, docs/control-protocol.md and
docs/target-app-parity.md.
The params text of select.same.symbolInstance, docs/mcp.md and
docs/target-app-parity.md now say that the object compared with can be
of any kind (a group member selected on its own, or an object that is
not a symbol instance, which leaves the selection empty), and that new
art on a locked artboard is not passed over while it stays selected.
A test covers the three cases.
The Global Edit Options dropdowns in the Control bar and in the
Properties panel's Quick Actions turn their button off while Global
Edit is on, but a popup that was already open stayed open and still
changed the options. Both dropdowns now close their popup while Global
Edit is on, and it stays closed after Global Edit stops until the
button is clicked again.
A selection that mixed kerning modes left the Kerning dropdown without
its Auto and Optical entries, so one mode could not be set on all of
it. The menu now lists both entries with neither checked.
The first UI pass applies the saved preferences again (#4837) by removing
the record of what was last applied. A Render Threads change made by a
control request between an earlier tick and that pass was then applied
with no previous value to compare, so the canvas renderer kept its old
thread count. The pass now forces the theme refresh and keeps the
record, so the thread change still restarts the renderer.
Done (pattern editing), Stop Editing Opacity Mask and Exit Isolation
Mode end the Type tool's edit first, so point type a click placed and
left empty is discarded. While a Cmd/Ctrl press had lent a selection
tool, the Type tool was not the active tool, its edit was not ended,
and the empty point type was saved into the pattern tile or the mask.
The lent tool is now given back first. A drag the lent tool started is
kept as its own step, with its last position applied when Real-time
Drawing and Editing is off.
A press or release of Backtick, or a window focus loss, that came after
the mouse button went up in the same frame changed whether the drag
committed a pattern-only transform. The release now takes Backtick's
state at the button-up from the frame's ordered input events.
optical_pair_after_rtl_separate_mark_uses_base_outline sets Noto Sans
Arabic, which is not in assets/fonts. Without CRAFT_FONTS_DIR the font
database returns a fallback face, so the test checked that face instead.
The test now skips when Noto Sans Arabic is missing, as the Shippori
Mincho tests in the same file do.
The marquee's segment test built a two-anchor path for every segment and ran
hit::intersects_rect on it. It now skips a segment whose anchors and handles
all lie past one side of the marquee, and tests the rest directly with the same
bounds and edge checks, so it selects the same segments.
A marquee over 10,000 anchors, press to release (anchor_bench, median of 7):
one open path of 200,000 anchors 84.9 ms -> 24.1 ms, 2,000 closed paths of 50
anchors 27.5 ms -> 13.5 ms, one closed path of 50,000 anchors 14.6 ms -> 8.5 ms.
A layer nested in a group whose art was all locked counted for nothing,
so the group was hidden or locked with the layer inside it. The selected
layer now counts as selected itself.
While Lock Guides was on, a guide path that was locked itself, or sat in
a locked object, was not kept, so the group around it was hidden. Guide
paths inside locked objects are now kept too.
A plug-in object that names a brush kind but ends before the brush's
name was dropped without a trace. When it was the file's only brush
object, the document kept the default brushes and showed no note. It is
now counted with the other brushes that can't be read, so the file gets
the Brushes panel note and no default brushes.
Holding Escape to cancel a floating Tools panel move or a ruler-origin
drag no longer passes its key repeats to the active tool or to the
other Escape actions (pattern edit, isolation, a flyout) once the
cancel frame has passed. The repeats are skipped until Escape is
released or pressed again. Tests hold Escape through both cancels in
each screen mode.
render_node_with and render_node_thumbnail_with drew into a new context
without setting its aliasing threshold, so AntiAlias::None still gave
smooth edges. Both now set the threshold from the options, as a page
render does.
On a multithreaded context, a drop shadow or glow drawn through a filter
layer is drawn into an offscreen that covers the art plus the blur's
spread, but its frame kept the view's visible rect. Art outside the view
but within the spread was culled, so it cast no shadow, and a long stroke
was outlined only near the view, so its far parts no longer fed the blur.
The offscreen's frame now takes its visible rect from its own area, and
the frame of content kept from an earlier split takes the window it is
drawn over.
The renderer outlined the whole stroke of an edited path on every frame,
so dragging one anchor of a path with many anchors re-outlined every
segment. A stroke line of more than 1,024 elements without dashes or a
width profile is now outlined only for the segments within the stroke's
reach of the view, inflated by the view's own size on each side. The
cached outline keeps that region, and a frame outside it outlines the
stroke again. Closed subpaths that a wide stroke crosses over stay whole,
and a cut closed subpath is walked from a removed segment so no cut lands
where it starts and ends. The pixels are the same.
At 1600 x 1200, the frame after one anchor moves takes 28 to 30 ms
instead of 103 ms on an open path of 200,000 anchors, and 10 ms instead
of 30 ms on a closed path of 50,000 anchors (the new anchor_bench
example: cargo run --release -p vectorcraft-engine --example anchor_bench).
`window.panel` on a panel whose group was hidden returned the whole
layout and every floating panel in every window. It now returns
`{floating}` with the panels of the restored panel's window, as it
does when it shows a panel that already floats.
Remove Text Tweaker's own start and end check: range_of already refuses
a bad offset given with an id, and like the other type commands Text
Tweaker now ignores start and end without one. The test's "2" case
gets an id so it still reaches the check, and docs/mcp.md lists
addon.type.textTweaker with the commands that refuse such offsets.
text.startList read a negative or fractional `at` as the end of the
text, so it could remove a typed list marker there and start a list.
The `start` and `end` of text.editRange, text.getRange,
text.setRangeStyle and the type shortcut commands given an `id` read
such values as the start or end of the text.
These now return "`at` must be a whole number" (naming the key) and
change nothing. Omitted or null keys keep their defaults, and 2.0 is
read as 2.
The Leading check now matches each word of the Leading field as a whole
word in languages that separate words with spaces, so a longer word such
as "liderlik" no longer stands in for "lider". Chinese, Japanese and
other unspaced scripts keep substring matching.
The Info row check now looks for the whole Tracking name in the label in
front of {tracking}, after taking out a Kerning or Leading name that
contains it. A word of the name found in the Size label, or a Kerning
label that contains the Tracking name, no longer passes.
The Czech Auto Leading row uses the singular "proklad", as the Leading
field does.
eframe's web runner reports a Cmd+V press both as the key press and, with the
clipboard's text, as a Paste event; the app pasted on each (egui's window
integration sends only the Paste). The web shell now drops the paste key's
press when a Paste is in the same frame's input; its release stays for
PasteChord, and a Cmd+V with no text on the clipboard keeps its press.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The test bound a port, freed it and then connected to it as a gone app.
Tests run in parallel, so another test's listener could take that port
meanwhile and accept this connection in place of its own client's,
whose connection was then reset when the listener closed: the cause of
the intermittent tcp_connection_is_closed_after_junk failures on
Windows. The gone app is now a listener that closes without a greeting.
While Select > Start Global Edit can run, the Control bar shows a Global Edit
Options dropdown after Recolor Artwork. It opens the Properties panel's Global
Edit Options (Match: Appearance, Size; On Artboards; Include Objects On
Canvas; Result(s)), so both set the same options, and it is off while Global
Edit is on, as the Properties dropdown is. The bar leaves it out when no room
is left for it.
Public reference:
https://helpx.adobe.com/illustrator/desktop/manage-objects/edit-objects/edit-similar-objects.html
The dropdown takes room ahead of the inline X/Y/W/H fields, so the ruler-origin
tests that read X/Y from the Control bar draw it 1700 px wide.
Opening a .ai or a PDF read and checked the whole file up to five
times: for the document it carries, for whether it holds the editor's
private data, for the import, for the editor's copy and for its File
Info. Each pass parses the file and walks every object, about 190 ms on
a 15 MB .ai.
- vectorcraft_pdf::open_pdf reads and checks the file once, on the
import's deep-stack thread, and import_opened imports pages from that
reading. document.open takes the carried document, the page count,
the editor's copy and the File Info from it. The editor's copy is read
after the page import and only when the file is not restored from a
carried document; so is the File Info.
- A .ai saved without PDF compatibility opened with pages that name
every page reads its editor's copy from the same reading.
- document.pdfInfo reads the file once for its pages and the thumbnail.
- The Open and Place dialogs read the file once to decide whether to
ask for pages, instead of twice.
A file whose PDF part fails the check still reads its editing data
separately, as before.
A blend's record in the editing data holds its Blend Options spacing.
Word 1 is the spacing: 0 is Smooth Color, 1 is Specified Steps and 2 is
Specified Distance. For Specified Distance, word 2 is the distance in
points as a 32-bit float. The reader took only records whose word 1 is
1, so every Smooth Color and Specified Distance blend came in as its
drawn steps.
A Smooth Color or Specified Distance blend now opens live with that
spacing where VectorCraft lays out as many steps between each pair of
keys as the file draws, in the same places and colors. Otherwise it
keeps its drawn steps with the existing note. The distance is read as
the shortest decimal that reads back as the stored float. The steps are
counted before anything is laid out, and a blend takes from the
import's budget only when its count matches. Specified Steps blends
are read as before, with the steps counted from the drawn ones.
Tests: import::ai::tests::a_blend_keeps_its_spacing,
a_blend_of_three_keys_keeps_its_distance_between_each_pair,
a_blend_whose_spacing_doesnt_lay_out_its_steps_keeps_them,
import::ai::blend::tests::records_say_their_spacing,
a_blend_takes_from_the_budget_only_when_it_is_laid_out,
a_distance_past_the_keys_lays_out_no_steps, and the fuzz case
ai_blend_spacing_never_panics.
A variable font can leave an ssNN feature empty in GSUB or GPOS and
supply its lookups through a FeatureVariations substitution. The
OpenType panel skipped those sets because the default feature has no
lookups. A set is now listed when a substitution in a variation record
replaces its feature with one that has lookups, and its name is read
from the default feature or the substituted one.
Records after one without conditions are ignored, and at most 4096
variation records and 4096 substitutions are read per table.
Follows #5447.
The room estimate for a large PNG read a row at a time counted 32 bytes
a column. The png reader also keeps at least 128 KiB of rows already
read, the 32 KiB deflate window and a growth step in its stream buffer,
and up to four rows for wide images, so a decode could be approved with
less room than reading takes. The estimate now comes from the PNG
header's row length and these buffers, and a PNG is read a row at a
time only while that estimate fits the decode allowance.
dotCooCoo